implementation of tantivy datastore + updated RPC calls to deal with tantivy + docs

Cargo.toml

@@ -1,8 +1,8 @@
 [package]
 name = "herodb"
 version = "0.0.1"
-authors = ["Pin Fang <fpfangpin@hotmail.com>"]
-edition = "2021"
+authors = ["ThreeFold Tech NV"]
+edition = "2024"
 
 [dependencies]
 anyhow = "1.0.59"
@@ -23,19 +23,10 @@ sha2 = "0.10"
 age = "0.10"
 secrecy = "0.8"
 ed25519-dalek = "2"
+x25519-dalek = "2"
 base64 = "0.22"
-# Lance vector database dependencies
-lance = "0.33"
-lance-index = "0.33"
-lance-linalg = "0.33"
-# Use Arrow version compatible with Lance 0.33
-arrow = "55.2"
-arrow-array = "55.2"
-arrow-schema = "55.2"
-parquet = "55.2"
-uuid = { version = "1.10", features = ["v4"] }
-reqwest = { version = "0.11", features = ["json"] }
-image = "0.25"
+jsonrpsee = { version = "0.26.0", features = ["http-client", "ws-client", "server", "macros"] }
+tantivy = "0.25.0"
 
 [dev-dependencies]
 redis = { version = "0.24", features = ["aio", "tokio-comp"] }

README.md

@@ -17,6 +17,8 @@ The main purpose of HeroDB is to offer a lightweight, embeddable, and Redis-comp
 - **Expiration**: Time-to-live (TTL) functionality for keys.
 - **Scanning**: Cursor-based iteration for keys and hash fields (`SCAN`, `HSCAN`).
 - **AGE Cryptography Commands**: HeroDB-specific extensions for cryptographic operations.
+- **Symmetric Encryption**: Stateless symmetric encryption using XChaCha20-Poly1305.
+- **Admin Database 0**: Centralized control for database management, access control, and per-database encryption.
 
 ## Quick Start
 
@@ -30,31 +32,14 @@ cargo build --release
 
 ### Running HeroDB
 
-You can start HeroDB with different backends and encryption options:
-
-#### Default `redb` Backend
+Launch HeroDB with the required `--admin-secret` flag, which encrypts the admin database (DB 0) and authorizes admin access. Optional flags include `--dir` for the database directory, `--port` for the TCP port (default 6379), `--sled` for the sled backend, and `--enable-rpc` to start the JSON-RPC management server on port 8080.
 
+Example:
 ```bash
-./target/release/herodb --dir /tmp/herodb_redb --port 6379
+./target/release/herodb --dir /tmp/herodb --admin-secret myadminsecret --port 6379 --enable-rpc
 ```
 
-#### `sled` Backend
-
-```bash
-./target/release/herodb --dir /tmp/herodb_sled --port 6379 --sled
-```
-
-#### `redb` with Encryption
-
-```bash
-./target/release/herodb --dir /tmp/herodb_encrypted --port 6379 --encrypt --key mysecretkey
-```
-
-#### `sled` with Encryption
-
-```bash
-./target/release/herodb --dir /tmp/herodb_sled_encrypted --port 6379 --sled --encrypt --key mysecretkey
-```
+For detailed launch options, see [Basics](docs/basics.md).
 
 ## Usage with Redis Clients
 
@@ -76,10 +61,24 @@ redis-cli -p 6379 SCAN 0 MATCH user:* COUNT 10
 #    2) 1) "user:1"
 ```
 
+## Cryptography
+
+HeroDB supports asymmetric encryption/signatures via AGE commands (X25519 for encryption, Ed25519 for signatures) in stateless or key-managed modes, and symmetric encryption via SYM commands. Keys are persisted in the admin database (DB 0) for managed modes.
+
+For details, see [AGE Cryptography](docs/age.md) and [Basics](docs/basics.md).
+
+## Database Management
+
+Databases are managed via JSON-RPC API, with metadata stored in the encrypted admin database (DB 0). Databases are public by default upon creation; use RPC to set them private, requiring access keys for SELECT operations (read or readwrite based on permissions). This includes per-database encryption keys, access control, and lifecycle management.
+
+For examples, see [JSON-RPC Examples](docs/rpc_examples.md) and [Admin DB 0 Model](docs/admin.md).
+
 ## Documentation
 
 For more detailed information on commands, features, and advanced usage, please refer to the documentation:
 
 - [Basics](docs/basics.md)
 - [Supported Commands](docs/cmds.md)
-- [AGE Cryptography](docs/age.md)
+- [AGE Cryptography](docs/age.md)
+- [Admin DB 0 Model (access control, per-db encryption)](docs/admin.md)
+- [JSON-RPC Examples (management API)](docs/rpc_examples.md)

docs/admin.md

@@ -0,0 +1,181 @@
+# Admin Database 0 (`0.db`)
+
+This page explains what the Admin Database `DB 0` is, why HeroDB uses it, and how to work with it as a developer and end-user. It’s a practical guide covering how databases are created, listed, secured with access keys, and encrypted using per-database secrets.
+
+## What is `DB 0`?
+
+`DB 0` is the control-plane for a HeroDB instance. It stores metadata for all user databases (`db_id >= 1`) so the server can:
+- Know which databases exist (without scanning the filesystem)
+- Enforce access control (public/private with access keys)
+- Enforce per-database encryption (whether a given database must be opened encrypted and with which write-only key)
+
+`DB 0` itself is always encrypted with the admin secret (the process-level secret provided at startup).
+
+## How `DB 0` is created and secured
+
+- `DB 0` lives at `<base_dir>/0.db`
+- It is always encrypted using the `admin secret` provided at process startup (using the `--admin-secret <secret>` CLI flag)
+- Only clients that provide the correct admin secret can `SELECT 0` (see “`SELECT` + `KEY`” below)
+
+At startup, the server bootstraps `DB 0` (initializes counters and structures) if it’s missing.
+
+## Metadata stored in `DB 0`
+
+Keys in `DB 0` (internal layout, but useful to understand how things work):
+
+- `admin:next_id`
+  - String counter holding the next id to allocate (initialized to `"1"`)
+
+- `admin:dbs`
+  - A hash acting as a set of existing database ids
+  - field = id (as string), value = `"1"`
+
+- `meta:db:<id>`
+  - A hash holding db-level metadata
+  - field `public` = `"true"` or `"false"` (defaults to `true` if missing)
+
+- `meta:db:<id>:keys`
+  - A hash mapping access-key hashes to the string `Permission:created_at_seconds`
+  - Examples: `Read:1713456789` or `ReadWrite:1713456789`
+  - The plaintext access keys are never stored; only their `SHA-256` hashes are kept
+
+- `meta:db:<id>:enc`
+   - A string holding the per-database encryption key used to open `<id>.db` encrypted
+   - This value is write-only from the perspective of the management APIs (it’s set at creation and never returned)
+
+- `age:key:<name>`
+   - Base64-encoded X25519 recipient (public encryption key) for named AGE keys
+- `age:privkey:<name>`
+   - Base64-encoded X25519 identity (secret encryption key) for named AGE keys
+- `age:signpub:<name>`
+   - Base64-encoded Ed25519 verify public key for named AGE keys
+- `age:signpriv:<name>`
+   - Base64-encoded Ed25519 signing secret key for named AGE keys
+
+> You don’t need to manipulate these keys directly; they’re listed to clarify the model. AGE keys are managed via AGE commands.
+
+## Database lifecycle
+
+1) Create a database (via JSON-RPC)
+- The server allocates an id from `admin:next_id`, registers it in `admin:dbs`, and defaults the database to `public=true`
+- If you pass an optional `encryption_key` during creation, the server persists it in `meta:db:<id>:enc`. That database will be opened in encrypted mode from then on
+
+2) Open and use a database
+- Clients select a database over RESP using `SELECT`
+- Authorization and encryption state are enforced using `DB 0` metadata
+
+3) Delete database files
+- Removing `<id>.db` removes the physical storage
+- `DB 0` remains the source of truth for existence and may be updated by future management methods as the system evolves
+
+## Access control model
+
+- Public database (default)
+  - Anyone can `SELECT <id>` with no key, and will get `ReadWrite` permission
+- Private database
+  - You must provide an access key when selecting the database
+  - The server hashes the provided key with `SHA-256` and checks membership in `meta:db:<id>:keys`
+  - Permissions are `Read` or `ReadWrite` depending on how the key was added
+- Admin `DB 0`
+  - Requires the exact admin secret as the `KEY` argument to `SELECT 0`
+  - Permission is `ReadWrite` when the secret matches
+
+### How to select databases with optional `KEY`
+
+- Public DB (no key required)
+  - `SELECT <id>`
+
+- Private DB (access key required)
+  - `SELECT <id> KEY <plaintext_key>`
+
+- Admin `DB 0` (admin secret required)
+  - `SELECT 0 KEY <admin_secret>`
+
+Examples (using `redis-cli`):
+```bash
+# Public database
+redis-cli -p $PORT SELECT 1
+# → OK
+
+# Private database
+redis-cli -p $PORT SELECT 2 KEY my-db2-access-key
+# → OK
+
+# Admin DB 0
+redis-cli -p $PORT SELECT 0 KEY my-admin-secret
+# → OK
+```
+
+## Per-database encryption
+
+- At database creation, you can provide an optional per-db encryption key
+- If provided, the server persists that key in `DB 0` as `meta:db:<id>:enc`
+- When you later open the database, the engine checks whether `meta:db:<id>:enc` exists to decide if it must open `<id>.db` in encrypted mode
+- The per-db key is not returned by RPC—it is considered write-only configuration data
+
+Operationally:
+- Create with encryption: pass a non-null `encryption_key` to the `createDatabase` RPC
+- Open later: simply `SELECT` the database; encryption is transparent to clients
+
+## Management via JSON-RPC
+
+You can manage databases using the management RPC (namespaced `herodb.*`). Typical operations:
+- `createDatabase(backend, config, encryption_key?)`
+  - Allocates a new id, sets optional encryption key
+- `listDatabases()`
+  - Lists database ids and info (including whether storage is currently encrypted)
+- `getDatabaseInfo(db_id)`
+  - Returns details: backend, encrypted flag, size on disk, `key_count`, timestamps, etc.
+- `addAccessKey(db_id, key, permissions)`
+  - Adds a `Read` or `ReadWrite` access key (permissions = `"read"` | `"readwrite"`)
+- `listAccessKeys(db_id)`
+  - Returns hashes and permissions; you can use these hashes to delete keys
+- `deleteAccessKey(db_id, key_hash)`
+  - Removes a key by its hash
+- `setDatabasePublic(db_id, public)`
+  - Toggles public/private
+
+Copyable JSON examples are provided in the [RPC examples documentation](./rpc_examples.md).
+
+## Typical flows
+
+1) Public, unencrypted database
+- Create a new database without an encryption key
+- Clients can immediately `SELECT <id>` without a key
+- You can later make it private and add keys if needed
+
+2) Private, encrypted database
+- Create passing an `encryption_key`
+- Mark it private (`setDatabasePublic false`) and add access keys
+- Clients must use `SELECT <id> KEY <plaintext_access_key>`
+- Storage opens in encrypted mode automatically
+
+## Security notes
+
+- Only `SHA-256` hashes of access keys are stored in `DB 0`; keep plaintext keys safe on the client side
+- The per-db encryption key is never exposed via the API after it is set
+- The admin secret must be kept secure; anyone with it can `SELECT 0` and perform administrative actions
+
+## Troubleshooting
+
+- `ERR invalid access key` when selecting a private db
+  - Ensure you passed the `KEY` argument: `SELECT <id> KEY <plaintext_key>`
+  - If you recently added the key, confirm the permissions and that you used the exact plaintext (hash must match)
+
+- `Database X not found`
+  - The id isn’t registered in `DB 0` (`admin:dbs`). Use the management APIs to create or list databases
+
+- Cannot `SELECT 0`
+  - The `KEY` must be the exact admin secret passed at server startup
+
+## Reference
+
+- Admin metadata lives in `DB 0` (`0.db`) and controls:
+  - Existence: `admin:dbs`
+  - Access: `meta:db:<id>.public` and `meta:db:<id>:keys`
+  - Encryption: `meta:db:<id>:enc`
+
+For command examples and management payloads:
+- RESP command basics: `docs/basics.md`
+- Supported commands: `docs/cmds.md`
+- JSON-RPC examples: `docs/rpc_examples.md`

docs/age.md

@@ -1,188 +1,96 @@
-# HeroDB AGE usage: Stateless vs Key‑Managed
+# HeroDB AGE Cryptography
 
-This document explains how to use the AGE cryptography commands exposed by HeroDB over the Redis protocol in two modes:
-- Stateless (ephemeral keys; nothing stored on the server)
-- Key‑managed (server‑persisted, named keys)
+HeroDB provides AGE-based asymmetric encryption and digital signatures over the Redis protocol using X25519 for encryption and Ed25519 for signatures. Keys can be used in stateless (ephemeral) or key-managed (persistent, named) modes.
 
-If you are new to the codebase, the exact tests that exercise these behaviors are:
-- [rust.test_07_age_stateless_suite()](herodb/tests/usage_suite.rs:495)
-- [rust.test_08_age_persistent_named_suite()](herodb/tests/usage_suite.rs:555)
+In key-managed mode, HeroDB uses a unified keypair concept: a single Ed25519 signing key is deterministically derived into X25519 keys for encryption, allowing one keypair to handle both encryption and signatures transparently.
 
-Implementation entry points:
-- [herodb/src/age.rs](herodb/src/age.rs)
-- Dispatch from [herodb/src/cmd.rs](herodb/src/cmd.rs)
+## Cryptographic Algorithms
 
-Note: Database-at-rest encryption flags in the test harness are unrelated to AGE commands; those flags control storage-level encryption of DB files. See the harness near [rust.start_test_server()](herodb/tests/usage_suite.rs:10).
+### X25519 (Encryption)
+- Elliptic-curve Diffie-Hellman key exchange for symmetric key derivation.
+- Used for encrypting/decrypting messages.
 
-## Quick start
+### Ed25519 (Signatures)
+- EdDSA digital signatures for message authentication.
+- Used for signing/verifying messages.
 
-Assuming the server is running on localhost on some $PORT:
+### Key Derivation
+Ed25519 signing keys are deterministically converted to X25519 keys for encryption. This enables a single keypair to support both operations without additional keys. Derivation uses the Ed25519 secret scalar clamped for X25519.
+
+In named keypairs, Ed25519 keys are stored, and X25519 keys are derived on-demand and cached.
+
+## Stateless Mode (Ephemeral Keys)
+No server-side storage; keys are provided with each command.
+
+Available commands:
+- `AGE GENENC`: Generate ephemeral X25519 keypair. Returns `[recipient, identity]`.
+- `AGE GENSIGN`: Generate ephemeral Ed25519 keypair. Returns `[verify_pub, sign_secret]`.
+- `AGE ENCRYPT <recipient> <message>`: Encrypt message. Returns base64 ciphertext.
+- `AGE DECRYPT <identity> <ciphertext_b64>`: Decrypt ciphertext. Returns plaintext.
+- `AGE SIGN <sign_secret> <message>`: Sign message. Returns base64 signature.
+- `AGE VERIFY <verify_pub> <message> <signature_b64>`: Verify signature. Returns 1 (valid) or 0 (invalid).
+
+Example:
 ```bash
-~/code/git.ourworld.tf/herocode/herodb/herodb/build.sh
-~/code/git.ourworld.tf/herocode/herodb/target/release/herodb --dir /tmp/data --debug --$PORT 6381 --encryption-key 1234 --encrypt
-```
+redis-cli AGE GENENC
+# → 1) "age1qz..."  # recipient (X25519 public)
+#    2) "AGE-SECRET-KEY-1..."  # identity (X25519 secret)
 
+redis-cli AGE ENCRYPT "age1qz..." "hello"
+# → base64_ciphertext
 
-```bash
-export PORT=6381
-# Generate an ephemeral keypair and encrypt/decrypt a message (stateless mode)
-redis-cli -p $PORT AGE GENENC
-# → returns an array: [recipient, identity]
-
-redis-cli -p $PORT AGE ENCRYPT <recipient> "hello world"
-# → returns ciphertext (base64 in a bulk string)
-
-redis-cli -p $PORT AGE DECRYPT <identity> <ciphertext_b64>
-# → returns "hello world"
-```
-
-For key‑managed mode, generate a named key once and reference it by name afterwards:
-
-```bash
-redis-cli -p $PORT AGE KEYGEN app1
-# → persists encryption keypair under name "app1"
-
-redis-cli -p $PORT AGE ENCRYPTNAME app1 "hello"
-redis-cli -p $PORT AGE DECRYPTNAME app1 <ciphertext_b64>
-```
-
-## Stateless AGE (ephemeral)
-
-Characteristics
-
-- No server‑side storage of keys.
-- You pass the actual key material with every call.
-- Not listable via AGE LIST.
-
-Commands and examples
-
-1) Ephemeral encryption keys
-
-```bash
-# Generate an ephemeral encryption keypair 
-redis-cli -p $PORT AGE GENENC
-# Example output (abridged):
-# 1) "age1qz..."          # recipient (public key) = can be used by others e.g. to verify what I sign
-# 2) "AGE-SECRET-KEY-1..." # identity (secret) = is like my private, cannot lose this one
-
-# Encrypt with the recipient public key
-redis-cli -p $PORT AGE ENCRYPT "age1qz..." "hello world"
-
-# → returns bulk string payload: base64 ciphertext (encrypted content)
-
-# Decrypt with the identity (secret) in other words your private key
-redis-cli -p $PORT AGE DECRYPT "AGE-SECRET-KEY-1..." "<ciphertext_b64>"
-# → "hello world"
-```
-
-2) Ephemeral signing keys
-
-> ? is this same as my private key
-
-```bash
-
-# Generate an ephemeral signing keypair
-redis-cli -p $PORT AGE GENSIGN
-# Example output:
-# 1) "<verify_pub_b64>"
-# 2) "<sign_secret_b64>"
-
-# Sign a message with the secret
-redis-cli -p $PORT AGE SIGN "<sign_secret_b64>" "msg"
-# → returns "<signature_b64>"
-
-# Verify with the public key
-redis-cli -p $PORT AGE VERIFY "<verify_pub_b64>" "msg" "<signature_b64>"
-# → 1 (valid) or 0 (invalid)
-```
-
-When to use
-- You do not want the server to store private keys.
-- You already manage key material on the client side.
-- You need ad‑hoc operations without persistence.
-
-Reference test: [rust.test_07_age_stateless_suite()](herodb/tests/usage_suite.rs:495)
-
-## Key‑managed AGE (persistent, named)
-
-Characteristics
-- Server generates and persists keypairs under a chosen name.
-- Clients refer to keys by name; raw secrets are not supplied on each call.
-- Keys are discoverable via AGE LIST.
-
-Commands and examples
-
-1) Named encryption keys
-
-```bash
-# Create/persist a named encryption keypair
-redis-cli -p $PORT AGE KEYGEN app1
-# → returns [recipient, identity] but also stores them under name "app1"
-
-> TODO: should not return identity (security, but there can be separate function to export it e.g. AGE EXPORTKEY app1)
-
-# Encrypt using the stored public key
-redis-cli -p $PORT AGE ENCRYPTNAME app1 "hello"
-# → returns bulk string payload: base64 ciphertext
-
-# Decrypt using the stored secret
-redis-cli -p $PORT AGE DECRYPTNAME app1 "<ciphertext_b64>"
+redis-cli AGE DECRYPT "AGE-SECRET-KEY-1..." base64_ciphertext
 # → "hello"
 ```
 
-2) Named signing keys
+## Key-Managed Mode (Persistent Named Keys)
+Keys are stored server-side under names. Supports unified keypairs for both encryption and signatures.
 
+Available commands:
+- `AGE KEYGEN <name>`: Generate and store unified keypair. Returns `[recipient, identity]` in age format.
+- `AGE SIGNKEYGEN <name>`: Generate and store Ed25519 signing keypair. Returns `[verify_pub, sign_secret]`.
+- `AGE ENCRYPTNAME <name> <message>`: Encrypt with named key. Returns base64 ciphertext.
+- `AGE DECRYPTNAME <name> <ciphertext_b64>`: Decrypt with named key. Returns plaintext.
+- `AGE SIGNNAME <name> <message>`: Sign with named key. Returns base64 signature.
+- `AGE VERIFYNAME <name> <message> <signature_b64>`: Verify with named key. Returns 1 or 0.
+- `AGE LIST`: List all stored key names. Returns sorted array of names.
+
+### AGE LIST Output
+Returns a flat, deduplicated, sorted array of key names (strings). Each name corresponds to a stored keypair, which may include encryption keys (X25519), signing keys (Ed25519), or both.
+
+Output format: `["name1", "name2", ...]`
+
+Example:
 ```bash
-# Create/persist a named signing keypair
-redis-cli -p $PORT AGE SIGNKEYGEN app1
-# → returns [verify_pub_b64, sign_secret_b64] and stores under name "app1"
-
-> TODO: should not return sign_secret_b64 (for security, but there can be separate function to export it e.g. AGE EXPORTSIGNKEY app1)
-
-# Sign using the stored secret
-redis-cli -p $PORT AGE SIGNNAME app1 "msg"
-# → returns "<signature_b64>"
-
-# Verify using the stored public key
-redis-cli -p $PORT AGE VERIFYNAME app1 "msg" "<signature_b64>"
-# → 1 (valid) or 0 (invalid)
+redis-cli AGE LIST
+# →  1) "<named_keypair_1>"
+#    2) "<named_keypair_2>"
 ```
 
-3) List stored AGE keys
+For unified keypairs (from `AGE KEYGEN`), the name handles both encryption (derived X25519) and signatures (stored Ed25519) transparently.
 
+Example with named keys:
 ```bash
-redis-cli -p $PORT AGE LIST
-# Example output includes labels such as "encpub" and your key names (e.g., "app1")
+redis-cli AGE KEYGEN app1
+# →  1) "age1..."  # recipient
+#    2) "AGE-SECRET-KEY-1..."  # identity
+
+redis-cli AGE ENCRYPTNAME app1 "secret message"
+# → base64_ciphertext
+
+redis-cli AGE DECRYPTNAME app1 base64_ciphertext
+# → "secret message"
+
+redis-cli AGE SIGNNAME app1 "message"
+# → base64_signature
+
+redis-cli AGE VERIFYNAME app1 "message" base64_signature
+# → 1
 ```
 
-When to use
-- You want centralized key storage/rotation and fewer secrets on the client.
-- You need names/labels for workflows and can trust the server with secrets.
-- You want discoverability (AGE LIST) and simpler client commands.
+## Choosing a Mode
+- **Stateless**: For ad-hoc operations without persistence; client manages keys.
+- **Key-managed**: For centralized key lifecycle; server stores keys for convenience and discoverability.
 
-Reference test: [rust.test_08_age_persistent_named_suite()](herodb/tests/usage_suite.rs:555)
-
-## Choosing a mode
-
-- Prefer Stateless when:
-  - Minimizing server trust for secret material is the priority.
-  - Clients already have a secure mechanism to store/distribute keys.
-- Prefer Key‑managed when:
-  - Centralized lifecycle, naming, and discoverability are beneficial.
-  - You plan to integrate rotation, ACLs, or auditability on the server side.
-
-## Security notes
-
-- Treat identities and signing secrets as sensitive; avoid logging them.
-- For key‑managed mode, ensure server storage (and backups) are protected.
-- AGE operations here are application‑level crypto and are distinct from database-at-rest encryption configured in the test harness.
-
-## Repository pointers
-
-- Stateless examples in tests: [rust.test_07_age_stateless_suite()](herodb/tests/usage_suite.rs:495)
-- Key‑managed examples in tests: [rust.test_08_age_persistent_named_suite()](herodb/tests/usage_suite.rs:555)
-- AGE implementation: [herodb/src/age.rs](herodb/src/age.rs)
-- Command dispatch: [herodb/src/cmd.rs](herodb/src/cmd.rs)
-- Bash demo: [herodb/examples/age_bash_demo.sh](herodb/examples/age_bash_demo.sh)
-- Rust persistent demo: [herodb/examples/age_persist_demo.rs](herodb/examples/age_persist_demo.rs)
-- Additional notes: [herodb/instructions/encrypt.md](herodb/instructions/encrypt.md)
+Implementation: [herodb/src/age.rs](herodb/src/age.rs) <br> 
+Tests: [herodb/tests/usage_suite.rs](herodb/tests/usage_suite.rs)

docs/basics.md

@@ -1,4 +1,58 @@
-Here's an expanded version of the cmds.md documentation to include the list commands:
+# HeroDB Basics
+
+## Launching HeroDB
+
+To launch HeroDB, use the binary with required and optional flags. The `--admin-secret` flag is mandatory, encrypting the admin database (DB 0) and authorizing admin access.
+
+### Launch Flags
+- `--dir <path>`: Directory for database files (default: current directory).
+- `--port <port>`: TCP port for Redis protocol (default: 6379).
+- `--debug`: Enable debug logging.
+- `--sled`: Use Sled backend (default: Redb).
+- `--enable-rpc`: Start JSON-RPC management server on port 8080.
+- `--rpc-port <port>`: Custom RPC port (default: 8080).
+- `--admin-secret <secret>`: Required secret for DB 0 encryption and admin access.
+
+Example:
+```bash
+./target/release/herodb --dir /tmp/herodb --admin-secret mysecret --port 6379 --enable-rpc
+```
+
+Deprecated flags (`--encrypt`, `--encryption-key`) are ignored for data DBs; per-database encryption is managed via RPC.
+
+## Admin Database (DB 0)
+
+DB 0 acts as the administrative database instance, storing metadata for all user databases (IDs >= 1). It controls existence, access control, and per-database encryption. DB 0 is always encrypted with the `--admin-secret`.
+
+When creating a new database, DB 0 allocates an ID, registers it, and optionally stores a per-database encryption key (write-only). Databases are public by default; use RPC to set them private, requiring access keys for SELECT (read or readwrite based on permissions). Keys are persisted in DB 0 for managed AGE operations.
+
+Access DB 0 with `SELECT 0 KEY <admin-secret>`.
+
+## Symmetric Encryption
+
+HeroDB supports stateless symmetric encryption via SYM commands, using XChaCha20-Poly1305 AEAD.
+
+Commands:
+- `SYM KEYGEN`: Generate 32-byte key. Returns base64-encoded key.
+- `SYM ENCRYPT <key_b64> <message>`: Encrypt message. Returns base64 ciphertext.
+- `SYM DECRYPT <key_b64> <ciphertext_b64>`: Decrypt. Returns plaintext.
+
+Example:
+```bash
+redis-cli SYM KEYGEN
+# → base64_key
+
+redis-cli SYM ENCRYPT base64_key "secret"
+# → base64_ciphertext
+
+redis-cli SYM DECRYPT base64_key base64_ciphertext
+# → "secret"
+```
+
+## RPC Options
+
+Enable the JSON-RPC server with `--enable-rpc` for database management. Methods include creating databases, managing access keys, and setting encryption. See [JSON-RPC Examples](./rpc_examples.md) for payloads.
+
 # HeroDB Commands
 
 HeroDB implements a subset of Redis commands over the Redis protocol. This document describes the available commands and their usage.
@@ -575,6 +629,29 @@ redis-cli -p $PORT AGE LIST
 #    2) "keyname2"
 ```
 
+## SYM Commands
+
+### SYM KEYGEN
+Generate a symmetric encryption key.
+```bash
+redis-cli -p $PORT SYM KEYGEN
+# → base64_encoded_32byte_key
+```
+
+### SYM ENCRYPT
+Encrypt a message with a symmetric key.
+```bash
+redis-cli -p $PORT SYM ENCRYPT <key_b64> "message"
+# → base64_encoded_ciphertext
+```
+
+### SYM DECRYPT
+Decrypt a ciphertext with a symmetric key.
+```bash
+redis-cli -p $PORT SYM DECRYPT <key_b64> <ciphertext_b64>
+# → decrypted_message
+```
+
 ## Server Information Commands
 
 ### INFO
@@ -621,3 +698,27 @@ This expanded documentation includes all the list commands that were implemented
 10. LINDEX - get element by index
 11. LRANGE - get range of elements
 
+
+## Updated Database Selection and Access Keys
+
+HeroDB uses an `Admin DB 0` to control database existence, access, and encryption. Access to data DBs can be public (no key) or private (requires a key). See detailed model in `docs/admin.md`.
+
+Examples:
+
+```bash
+# Public database (no key required)
+redis-cli -p $PORT SELECT 1
+# → OK
+```
+
+```bash
+# Private database (requires access key)
+redis-cli -p $PORT SELECT 2 KEY my-db2-access-key
+# → OK
+```
+
+```bash
+# Admin DB 0 (requires admin secret)
+redis-cli -p $PORT SELECT 0 KEY my-admin-secret
+# → OK
+```

docs/cmds.md

@@ -122,4 +122,27 @@ redis-cli -p 6379 --rdb dump.rdb
 
 # Import to sled
 redis-cli -p 6381 --pipe < dump.rdb
+```
+
+## Authentication and Database Selection
+
+HeroDB uses an `Admin DB 0` to govern database existence, access and per-db encryption. Access control is enforced via `Admin DB 0` metadata. See the full model in `docs/admin.md`.
+
+Examples:
+```bash
+# Public database (no key required)
+redis-cli -p $PORT SELECT 1
+# → OK
+```
+
+```bash
+# Private database (requires access key)
+redis-cli -p $PORT SELECT 2 KEY my-db2-access-key
+# → OK
+```
+
+```bash
+# Admin DB 0 (requires admin secret)
+redis-cli -p $PORT SELECT 0 KEY my-admin-secret
+# → OK
 ```

docs/lance_vector_db.md

@@ -1,454 +0,0 @@
-# Lance Vector Database Operations
-
-HeroDB includes a powerful vector database integration using Lance, enabling high-performance vector storage, search, and multimodal data management. By default, it uses Ollama for local text embeddings, with support for custom external embedding services.
-
-## Overview
-
-The Lance vector database integration provides:
-
-- **High-performance vector storage** using Lance's columnar format
-- **Local Ollama integration** for text embeddings (default, no external dependencies)
-- **Custom embedding service support** for advanced use cases
-- **Text embedding support** (images via custom services)
-- **Vector similarity search** with configurable parameters
-- **Scalable indexing** with IVF_PQ (Inverted File with Product Quantization)
-- **Redis-compatible command interface**
-
-## Architecture
-
-```
-┌─────────────────┐    ┌──────────────────┐    ┌─────────────────┐
-│   HeroDB        │    │  External        │    │   Lance         │
-│   Redis Server  │◄──►│  Embedding       │    │   Vector Store  │
-│                 │    │  Service         │    │                 │
-└─────────────────┘    └──────────────────┘    └─────────────────┘
-         │                       │                       │
-         │                       │                       │
-    Redis Protocol          HTTP API              Arrow/Parquet
-    Commands                JSON Requests         Columnar Storage
-```
-
-### Key Components
-
-1. **Lance Store**: High-performance columnar vector storage
-2. **Ollama Integration**: Local embedding service (default)
-3. **Custom Embedding Service**: Optional HTTP API for advanced use cases
-4. **Redis Command Interface**: Familiar Redis-style commands
-5. **Arrow Schema**: Flexible schema definition for metadata
-
-## Configuration
-
-### Default Setup (Ollama)
-
-HeroDB uses Ollama by default for text embeddings. No configuration is required if Ollama is running locally:
-
-```bash
-# Install Ollama (if not already installed)
-# Visit: https://ollama.ai
-
-# Pull the embedding model
-ollama pull nomic-embed-text
-
-# Ollama automatically runs on localhost:11434
-# HeroDB will use this by default
-```
-
-**Default Configuration:**
-- **URL**: `http://localhost:11434`
-- **Model**: `nomic-embed-text`
-- **Dimensions**: 768 (for nomic-embed-text)
-
-### Custom Embedding Service (Optional)
-
-To use a custom embedding service instead of Ollama:
-
-```bash
-# Set custom embedding service URL
-redis-cli HSET config:core:aiembed url "http://your-embedding-service:8080/embed"
-
-# Optional: Set authentication if required
-redis-cli HSET config:core:aiembed token "your-api-token"
-```
-
-### Embedding Service API Contracts
-
-#### Ollama API (Default)
-HeroDB calls Ollama using this format:
-
-```bash
-POST http://localhost:11434/api/embeddings
-Content-Type: application/json
-
-{
-  "model": "nomic-embed-text",
-  "prompt": "Your text to embed"
-}
-```
-
-Response:
-```json
-{
-  "embedding": [0.1, 0.2, 0.3, ...]
-}
-```
-
-#### Custom Service API
-Your custom embedding service should accept POST requests with this JSON format:
-
-```json
-{
-  "texts": ["text1", "text2"],           // Optional: array of texts
-  "images": ["base64_image1", "base64_image2"], // Optional: base64 encoded images
-  "model": "your-model-name"             // Optional: model specification
-}
-```
-
-And return responses in this format:
-
-```json
-{
-  "embeddings": [[0.1, 0.2, ...], [0.3, 0.4, ...]],  // Array of embedding vectors
-  "model": "model-name",                               // Model used
-  "usage": {                                          // Optional usage stats
-    "tokens": 100,
-    "requests": 2
-  }
-}
-```
-
-## Commands Reference
-
-### Dataset Management
-
-#### LANCE CREATE
-Create a new vector dataset with specified dimensions and optional schema.
-
-```bash
-LANCE CREATE <dataset> DIM <dimension> [SCHEMA field:type ...]
-```
-
-**Parameters:**
-- `dataset`: Name of the dataset
-- `dimension`: Vector dimension (e.g., 384, 768, 1536)
-- `field:type`: Optional metadata fields (string, int, float, bool)
-
-**Examples:**
-```bash
-# Create a simple dataset for 384-dimensional vectors
-LANCE CREATE documents DIM 384
-
-# Create dataset with metadata schema
-LANCE CREATE products DIM 768 SCHEMA category:string price:float available:bool
-```
-
-#### LANCE LIST
-List all available datasets.
-
-```bash
-LANCE LIST
-```
-
-**Returns:** Array of dataset names
-
-#### LANCE INFO
-Get information about a specific dataset.
-
-```bash
-LANCE INFO <dataset>
-```
-
-**Returns:** Dataset metadata including name, version, row count, and schema
-
-#### LANCE DROP
-Delete a dataset and all its data.
-
-```bash
-LANCE DROP <dataset>
-```
-
-### Data Operations
-
-#### LANCE STORE
-Store multimodal data (text/images) with automatic embedding generation.
-
-```bash
-LANCE STORE <dataset> [TEXT <text>] [IMAGE <base64>] [key value ...]
-```
-
-**Parameters:**
-- `dataset`: Target dataset name
-- `TEXT`: Text content to embed
-- `IMAGE`: Base64-encoded image to embed
-- `key value`: Metadata key-value pairs
-
-**Examples:**
-```bash
-# Store text with metadata
-LANCE STORE documents TEXT "Machine learning is transforming industries" category "AI" author "John Doe"
-
-# Store image with metadata
-LANCE STORE images IMAGE "iVBORw0KGgoAAAANSUhEUgAA..." category "nature" tags "landscape,mountains"
-
-# Store both text and image
-LANCE STORE multimodal TEXT "Beautiful sunset" IMAGE "base64data..." location "California"
-```
-
-**Returns:** Unique ID of the stored item
-
-### Search Operations
-
-#### LANCE SEARCH
-Search using a raw vector.
-
-```bash
-LANCE SEARCH <dataset> VECTOR <vector> K <k> [NPROBES <n>] [REFINE <r>]
-```
-
-**Parameters:**
-- `dataset`: Dataset to search
-- `vector`: Comma-separated vector values (e.g., "0.1,0.2,0.3")
-- `k`: Number of results to return
-- `NPROBES`: Number of partitions to search (optional)
-- `REFINE`: Refine factor for better accuracy (optional)
-
-**Example:**
-```bash
-LANCE SEARCH documents VECTOR "0.1,0.2,0.3,0.4" K 5 NPROBES 10
-```
-
-#### LANCE SEARCH.TEXT
-Search using text query (automatically embedded).
-
-```bash
-LANCE SEARCH.TEXT <dataset> <query_text> K <k> [NPROBES <n>] [REFINE <r>]
-```
-
-**Parameters:**
-- `dataset`: Dataset to search
-- `query_text`: Text query to search for
-- `k`: Number of results to return
-- `NPROBES`: Number of partitions to search (optional)
-- `REFINE`: Refine factor for better accuracy (optional)
-
-**Example:**
-```bash
-LANCE SEARCH.TEXT documents "artificial intelligence applications" K 10 NPROBES 20
-```
-
-**Returns:** Array of results with distance scores and metadata
-
-### Embedding Operations
-
-#### LANCE EMBED.TEXT
-Generate embeddings for text without storing.
-
-```bash
-LANCE EMBED.TEXT <text1> [text2] [text3] ...
-```
-
-**Example:**
-```bash
-LANCE EMBED.TEXT "Hello world" "Machine learning" "Vector database"
-```
-
-**Returns:** Array of embedding vectors
-
-### Index Management
-
-#### LANCE CREATE.INDEX
-Create a vector index for faster search performance.
-
-```bash
-LANCE CREATE.INDEX <dataset> <index_type> [PARTITIONS <n>] [SUBVECTORS <n>]
-```
-
-**Parameters:**
-- `dataset`: Dataset to index
-- `index_type`: Index type (currently supports "IVF_PQ")
-- `PARTITIONS`: Number of partitions (default: 256)
-- `SUBVECTORS`: Number of sub-vectors for PQ (default: 16)
-
-**Example:**
-```bash
-LANCE CREATE.INDEX documents IVF_PQ PARTITIONS 512 SUBVECTORS 32
-```
-
-## Usage Patterns
-
-### 1. Document Search System
-
-```bash
-# Setup
-LANCE CREATE documents DIM 384 SCHEMA title:string content:string category:string
-
-# Store documents
-LANCE STORE documents TEXT "Introduction to machine learning algorithms" title "ML Basics" category "education"
-LANCE STORE documents TEXT "Deep learning neural networks explained" title "Deep Learning" category "education"
-LANCE STORE documents TEXT "Building scalable web applications" title "Web Dev" category "programming"
-
-# Create index for better performance
-LANCE CREATE.INDEX documents IVF_PQ PARTITIONS 256
-
-# Search
-LANCE SEARCH.TEXT documents "neural networks" K 5
-```
-
-### 2. Image Similarity Search
-
-```bash
-# Setup
-LANCE CREATE images DIM 512 SCHEMA filename:string tags:string
-
-# Store images (base64 encoded)
-LANCE STORE images IMAGE "iVBORw0KGgoAAAANSUhEUgAA..." filename "sunset.jpg" tags "nature,landscape"
-LANCE STORE images IMAGE "iVBORw0KGgoAAAANSUhEUgBB..." filename "city.jpg" tags "urban,architecture"
-
-# Search by image
-LANCE STORE temp_search IMAGE "query_image_base64..."
-# Then use the returned ID to get embedding and search
-```
-
-### 3. Multimodal Content Management
-
-```bash
-# Setup
-LANCE CREATE content DIM 768 SCHEMA type:string source:string
-
-# Store mixed content
-LANCE STORE content TEXT "Product description for smartphone" type "product" source "catalog"
-LANCE STORE content IMAGE "product_image_base64..." type "product_image" source "catalog"
-
-# Search across all content types
-LANCE SEARCH.TEXT content "smartphone features" K 10
-```
-
-## Performance Considerations
-
-### Vector Dimensions
-- **384**: Good for general text (e.g., sentence-transformers)
-- **768**: Standard for BERT-like models
-- **1536**: OpenAI text-embedding-ada-002
-- **Higher dimensions**: Better accuracy but slower search
-
-### Index Configuration
-- **More partitions**: Better for larger datasets (>100K vectors)
-- **More sub-vectors**: Better compression but slower search
-- **NPROBES**: Higher values = better accuracy, slower search
-
-### Best Practices
-
-1. **Create indexes** for datasets with >1000 vectors
-2. **Use appropriate dimensions** based on your embedding model
-3. **Configure NPROBES** based on accuracy vs speed requirements
-4. **Batch operations** when possible for better performance
-5. **Monitor embedding service** response times and rate limits
-
-## Error Handling
-
-Common error scenarios and solutions:
-
-### Embedding Service Errors
-```bash
-# Error: Embedding service not configured
-ERR Embedding service URL not configured. Set it with: HSET config:core:aiembed url <YOUR_EMBEDDING_SERVICE_URL>
-
-# Error: Service unavailable
-ERR Embedding service returned error 404 Not Found
-```
-
-**Solution:** Ensure embedding service is running and URL is correct.
-
-### Dataset Errors
-```bash
-# Error: Dataset doesn't exist
-ERR Dataset 'mydata' does not exist
-
-# Error: Dimension mismatch
-ERR Vector dimension mismatch: expected 384, got 768
-```
-
-**Solution:** Create dataset first or check vector dimensions.
-
-### Search Errors
-```bash
-# Error: Invalid vector format
-ERR Invalid vector format
-
-# Error: No index available
-ERR No index available for fast search
-```
-
-**Solution:** Check vector format or create an index.
-
-## Integration Examples
-
-### With Python
-```python
-import redis
-import json
-
-r = redis.Redis(host='localhost', port=6379)
-
-# Create dataset
-r.execute_command('LANCE', 'CREATE', 'docs', 'DIM', '384')
-
-# Store document
-result = r.execute_command('LANCE', 'STORE', 'docs', 
-                          'TEXT', 'Machine learning tutorial',
-                          'category', 'education')
-print(f"Stored with ID: {result}")
-
-# Search
-results = r.execute_command('LANCE', 'SEARCH.TEXT', 'docs', 
-                           'machine learning', 'K', '5')
-print(f"Search results: {results}")
-```
-
-### With Node.js
-```javascript
-const redis = require('redis');
-const client = redis.createClient();
-
-// Create dataset
-await client.sendCommand(['LANCE', 'CREATE', 'docs', 'DIM', '384']);
-
-// Store document
-const id = await client.sendCommand(['LANCE', 'STORE', 'docs', 
-                                   'TEXT', 'Deep learning guide',
-                                   'category', 'AI']);
-
-// Search
-const results = await client.sendCommand(['LANCE', 'SEARCH.TEXT', 'docs',
-                                        'deep learning', 'K', '10']);
-```
-
-## Monitoring and Maintenance
-
-### Health Checks
-```bash
-# Check if Lance store is available
-LANCE LIST
-
-# Check dataset health
-LANCE INFO mydataset
-
-# Test embedding service
-LANCE EMBED.TEXT "test"
-```
-
-### Maintenance Operations
-```bash
-# Backup: Use standard Redis backup procedures
-# The Lance data is stored separately in the data directory
-
-# Cleanup: Remove unused datasets
-LANCE DROP old_dataset
-
-# Reindex: Drop and recreate indexes if needed
-LANCE DROP dataset_name
-LANCE CREATE dataset_name DIM 384
-# Re-import data
-LANCE CREATE.INDEX dataset_name IVF_PQ
-```
-
-This integration provides a powerful foundation for building AI-powered applications with vector search capabilities while maintaining the familiar Redis interface.

docs/rpc_examples.md

@@ -0,0 +1,141 @@
+# HeroDB JSON-RPC Examples
+
+These examples show full JSON-RPC 2.0 payloads for managing HeroDB via the RPC API (enable with `--enable-rpc`). Methods are named as `hero_<function>`. Params are positional arrays; enum values are strings (e.g., `"Redb"`). Copy-paste into Postman or similar clients.
+
+## Database Management
+
+### Create Database
+Creates a new database with optional per-database encryption key (stored write-only in Admin DB 0).
+
+```json
+{
+  "jsonrpc": "2.0",
+  "id": 1,
+  "method": "hero_createDatabase",
+  "params": [
+    "Redb",
+    { "name": null, "storage_path": null, "max_size": null, "redis_version": null },
+    null
+  ]
+}
+```
+
+With encryption:
+```json
+{
+  "jsonrpc": "2.0",
+  "id": 2,
+  "method": "hero_createDatabase",
+  "params": [
+    "Sled",
+    { "name": "secure-db", "storage_path": null, "max_size": null, "redis_version": null },
+    "my-per-db-encryption-key"
+  ]
+}
+```
+
+### List Databases
+Returns array of database infos (id, backend, encrypted status, size, etc.).
+
+```json
+{
+  "jsonrpc": "2.0",
+  "id": 3,
+  "method": "hero_listDatabases",
+  "params": []
+}
+```
+
+### Get Database Info
+Retrieves detailed info for a specific database.
+
+```json
+{
+  "jsonrpc": "2.0",
+  "id": 4,
+  "method": "hero_getDatabaseInfo",
+  "params": [1]
+}
+```
+
+### Delete Database
+Removes physical database file; metadata remains in Admin DB 0.
+
+```json
+{
+  "jsonrpc": "2.0",
+  "id": 5,
+  "method": "hero_deleteDatabase",
+  "params": [1]
+}
+```
+
+## Access Control
+
+### Add Access Key
+Adds a hashed access key for private databases. Permissions: `"read"` or `"readwrite"`.
+
+```json
+{
+  "jsonrpc": "2.0",
+  "id": 6,
+  "method": "hero_addAccessKey",
+  "params": [2, "my-access-key", "readwrite"]
+}
+```
+
+### List Access Keys
+Returns array of key hashes, permissions, and creation timestamps.
+
+```json
+{
+  "jsonrpc": "2.0",
+  "id": 7,
+  "method": "hero_listAccessKeys",
+  "params": [2]
+}
+```
+
+### Delete Access Key
+Removes key by its SHA-256 hash.
+
+```json
+{
+  "jsonrpc": "2.0",
+  "id": 8,
+  "method": "hero_deleteAccessKey",
+  "params": [2, "0123abcd...keyhash..."]
+}
+```
+
+### Set Database Public/Private
+Toggles public access (default true). Private databases require access keys.
+
+```json
+{
+  "jsonrpc": "2.0",
+  "id": 9,
+  "method": "hero_setDatabasePublic",
+  "params": [2, false]
+}
+```
+
+## Server Info
+
+### Get Server Stats
+Returns stats like total databases and uptime.
+
+```json
+{
+  "jsonrpc": "2.0",
+  "id": 10,
+  "method": "hero_getServerStats",
+  "params": []
+}
+```
+
+## Notes
+- Per-database encryption keys are write-only; set at creation and used transparently.
+- Access keys are hashed (SHA-256) for storage; provide plaintext in requests.
+- Backend options: `"Redb"` (default) or `"Sled"`.
+- Config object fields (name, storage_path, etc.) are optional and currently ignored but positional.

docs/tantivy.md

@@ -0,0 +1,253 @@
+# Tantivy Full‑Text Backend (JSON‑RPC)
+
+This document explains how to use HeroDB’s Tantivy-backed full‑text search as a dedicated database backend and provides copy‑pasteable JSON‑RPC requests. Tantivy is available only for non‑admin databases (db_id >= 1). Admin DB 0 always uses Redb/Sled and rejects FT operations.
+
+Important characteristics:
+- Tantivy is a third backend alongside Redb and Sled. It provides search indexes only; there is no KV store backing it.
+- On Tantivy databases, Redis KV/list/hash commands are rejected; only FT commands and basic control (SELECT, CLIENT, INFO, etc.) are allowed.
+- FT JSON‑RPC is namespaced as "herodb" and methods are named with underscore: herodb_ftCreate, herodb_ftAdd, herodb_ftSearch, herodb_ftDel, herodb_ftInfo, herodb_ftDrop.
+
+Reference to server implementation:
+- RPC methods are defined in [rust.trait Rpc()](src/rpc.rs:70):
+  - [rust.fn ft_create()](src/rpc.rs:121)
+  - [rust.fn ft_add()](src/rpc.rs:130)
+  - [rust.fn ft_search()](src/rpc.rs:141)
+  - [rust.fn ft_del()](src/rpc.rs:154)
+  - [rust.fn ft_info()](src/rpc.rs:158)
+  - [rust.fn ft_drop()](src/rpc.rs:162)
+
+Notes on responses:
+- ftCreate/ftAdd/ftDel/ftDrop return a JSON boolean: true on success.
+- ftSearch/ftInfo return a JSON object with a single key "resp" containing a RESP‑encoded string (wire format used by Redis). You can display or parse it on the client side as needed.
+
+RESP usage (redis-cli):
+- For RESP clients, you must SELECT the Tantivy database first. SELECT now succeeds for Tantivy DBs without opening KV storage.
+- After SELECT, you can run FT.* commands within that DB context.
+
+Example with redis-cli:
+```bash
+# Connect to server
+redis-cli -p 6379
+
+# Select Tantivy DB 1 (public by default)
+SELECT 1
+# → OK
+
+# Create index
+FT.CREATE product_catalog SCHEMA title TEXT description TEXT category TAG price NUMERIC rating NUMERIC location GEO
+# → OK
+
+# Add a document
+FT.ADD product_catalog product:1 1.0 title "Wireless Bluetooth Headphones" description "Premium noise-canceling headphones with 30-hour battery life" category "electronics,audio" price 299.99 rating 4.5 location "-122.4194,37.7749"
+# → OK
+
+# Search
+FT.SEARCH product_catalog wireless LIMIT 0 3
+# → RESP array with hits
+```
+
+Storage layout (on disk):
+- Indices are stored per database under:
+  - <base_dir>/search_indexes/<db_id>/<index_name>
+- Example: /tmp/test/search_indexes/1/product_catalog
+
+0) Create a new Tantivy database
+
+Use herodb_createDatabase with backend "Tantivy". DB 0 cannot be Tantivy.
+
+```json
+{
+  "jsonrpc": "2.0",
+  "id": 1,
+  "method": "herodb_createDatabase",
+  "params": [
+    "Tantivy",
+    { "name": "search-db", "storage_path": null, "max_size": null, "redis_version": null },
+    null
+  ]
+}
+```
+
+The response contains the allocated db_id (>= 1). Use that id in the calls below.
+
+1) FT.CREATE — create an index with schema
+
+Method: herodb_ftCreate → [rust.fn ft_create()](src/rpc.rs:121)
+
+Schema format is an array of tuples: [ [field_name, field_type, [options...] ], ... ]
+Supported field types: "TEXT", "NUMERIC" (defaults to F64), "TAG", "GEO"
+Supported options (subset): "WEIGHT", "SORTABLE", "NOINDEX", "SEPARATOR", "CASESENSITIVE"
+
+```json
+{
+  "jsonrpc": "2.0",
+  "id": 2,
+  "method": "herodb_ftCreate",
+  "params": [
+    1,
+    "product_catalog",
+    [
+      ["title", "TEXT", ["SORTABLE"]],
+      ["description", "TEXT", []],
+      ["category", "TAG", ["SEPARATOR", ","]],
+      ["price", "NUMERIC", ["SORTABLE"]],
+      ["rating", "NUMERIC", []],
+      ["location", "GEO", []]
+    ]
+  ]
+}
+```
+
+Returns: true on success.
+
+2) FT.ADD — add or replace a document
+
+Method: herodb_ftAdd → [rust.fn ft_add()](src/rpc.rs:130)
+
+Fields is an object (map) of field_name → value (all values are sent as strings). GEO expects "lat,lon".
+
+```json
+{
+  "jsonrpc": "2.0",
+  "id": 3,
+  "method": "herodb_ftAdd",
+  "params": [
+    1,
+    "product_catalog",
+    "product:1",
+    1.0,
+    {
+      "title": "Wireless Bluetooth Headphones",
+      "description": "Premium noise-canceling headphones with 30-hour battery life",
+      "category": "electronics,audio",
+      "price": "299.99",
+      "rating": "4.5",
+      "location": "-122.4194,37.7749"
+    }
+  ]
+}
+```
+
+Returns: true on success.
+
+3) FT.SEARCH — query an index
+
+Method: herodb_ftSearch → [rust.fn ft_search()](src/rpc.rs:141)
+
+Parameters: (db_id, index_name, query, filters?, limit?, offset?, return_fields?)
+- filters: array of [field, value] pairs (Equals filter)
+- limit/offset: numbers (defaults: limit=10, offset=0)
+- return_fields: array of field names to include (optional)
+
+Simple query:
+```json
+{
+  "jsonrpc": "2.0",
+  "id": 4,
+  "method": "herodb_ftSearch",
+  "params": [1, "product_catalog", "wireless", null, 10, 0, null]
+}
+```
+
+Pagination + filters + selected fields:
+```json
+{
+  "jsonrpc": "2.0",
+  "id": 5,
+  "method": "herodb_ftSearch",
+  "params": [
+    1,
+    "product_catalog",
+    "mouse",
+    [["category", "electronics"]],
+    5,
+    0,
+    ["title", "price", "rating"]
+  ]
+}
+```
+
+Response shape:
+```json
+{
+  "jsonrpc": "2.0",
+  "id": 5,
+  "result": { "resp": "*...RESP encoded array..." }
+}
+```
+
+4) FT.INFO — index metadata
+
+Method: herodb_ftInfo → [rust.fn ft_info()](src/rpc.rs:158)
+
+```json
+{
+  "jsonrpc": "2.0",
+  "id": 6,
+  "method": "herodb_ftInfo",
+  "params": [1, "product_catalog"]
+}
+```
+
+Response shape:
+```json
+{
+  "jsonrpc": "2.0",
+  "id": 6,
+  "result": { "resp": "*...RESP encoded array with fields and counts..." }
+}
+```
+
+5) FT.DEL — delete by doc id
+
+Method: herodb_ftDel → [rust.fn ft_del()](src/rpc.rs:154)
+
+```json
+{
+  "jsonrpc": "2.0",
+  "id": 7,
+  "method": "herodb_ftDel",
+  "params": [1, "product_catalog", "product:1"]
+}
+```
+
+Returns: true on success. Note: current implementation logs and returns success; physical delete may be a no‑op until delete is finalized in the engine.
+
+6) FT.DROP — drop an index
+
+Method: herodb_ftDrop → [rust.fn ft_drop()](src/rpc.rs:162)
+
+```json
+{
+  "jsonrpc": "2.0",
+  "id": 8,
+  "method": "herodb_ftDrop",
+  "params": [1, "product_catalog"]
+}
+```
+
+Returns: true on success.
+
+Field types and options
+
+- TEXT: stored/indexed/tokenized text. "SORTABLE" marks it fast (stored + fast path in our wrapper).
+- NUMERIC: stored/indexed numeric; default precision F64. "SORTABLE" enables fast column.
+- TAG: exact matching terms. Options: "SEPARATOR" (default ","), "CASESENSITIVE" (default false).
+- GEO: "lat,lon" string; stored as two numeric fields internally.
+
+Backend and permission gating
+
+- FT methods are rejected on DB 0.
+- FT methods require the database backend to be Tantivy; otherwise RPC returns an error.
+- Write‑like FT methods (create/add/del/drop) follow the same permission model as Redis writes on selected databases.
+
+Troubleshooting
+
+- "DB backend is not Tantivy": ensure the database was created with backend "Tantivy".
+- "FT not allowed on DB 0": use a non‑admin database id (>= 1).
+- Empty search results: confirm that the queried fields are tokenized/indexed (TEXT) and that documents were added successfully.
+
+Related docs
+
+- Command‑level search overview: [docs/search.md](docs/search.md:1)
+- RPC definitions: [src/rpc.rs](src/rpc.rs:1)

examples/README.md

@@ -1,191 +1,6 @@
-# HeroDB Examples
+# HeroDB Tantivy Search Examples
 
-This directory contains examples demonstrating HeroDB's capabilities including full-text search powered by Tantivy and vector database operations using Lance.
-
-## Available Examples
-
-1. **[Tantivy Search Demo](#tantivy-search-demo-bash-script)** - Full-text search capabilities
-2. **[Lance Vector Database Demo](#lance-vector-database-demo-bash-script)** - Vector database and AI operations
-3. **[AGE Encryption Demo](age_bash_demo.sh)** - Cryptographic operations
-4. **[Simple Demo](simple_demo.sh)** - Basic Redis operations
-
----
-
-## Lance Vector Database Demo (Bash Script)
-
-### Overview
-The `lance_vector_demo.sh` script provides a comprehensive demonstration of HeroDB's vector database capabilities using Lance. It showcases vector storage, similarity search, multimodal data handling, and AI-powered operations with external embedding services.
-
-### Prerequisites
-1. **HeroDB Server**: The server must be running (default port 6379)
-2. **Redis CLI**: The `redis-cli` tool must be installed and available in your PATH
-3. **Embedding Service** (optional): For full functionality, set up an external embedding service
-
-### Running the Demo
-
-#### Step 1: Start HeroDB Server
-```bash
-# From the project root directory
-cargo run -- --dir ./test_data --port 6379
-```
-
-#### Step 2: Run the Demo (in a new terminal)
-```bash
-# From the project root directory
-./examples/lance_vector_demo.sh
-```
-
-### What the Demo Covers
-
-The script demonstrates comprehensive vector database operations:
-
-1. **Dataset Management**
-   - Creating vector datasets with custom dimensions
-   - Defining schemas with metadata fields
-   - Listing and inspecting datasets
-   - Dataset information and statistics
-
-2. **Embedding Operations**
-   - Text embedding generation via external services
-   - Multimodal embedding support (text + images)
-   - Batch embedding operations
-
-3. **Data Storage**
-   - Storing text documents with automatic embedding
-   - Storing images with metadata
-   - Multimodal content storage
-   - Rich metadata support
-
-4. **Vector Search**
-   - Similarity search with raw vectors
-   - Text-based semantic search
-   - Configurable search parameters (K, NPROBES, REFINE)
-   - Cross-modal search capabilities
-
-5. **Index Management**
-   - Creating IVF_PQ indexes for performance
-   - Custom index parameters
-   - Performance optimization
-
-6. **Advanced Features**
-   - Error handling and recovery
-   - Performance testing concepts
-   - Monitoring and maintenance
-   - Cleanup operations
-
-### Key Lance Commands Demonstrated
-
-#### Dataset Management
-```bash
-# Create vector dataset
-LANCE CREATE documents DIM 384
-
-# Create dataset with schema
-LANCE CREATE products DIM 768 SCHEMA category:string price:float available:bool
-
-# List datasets
-LANCE LIST
-
-# Get dataset information
-LANCE INFO documents
-```
-
-#### Data Operations
-```bash
-# Store text with metadata
-LANCE STORE documents TEXT "Machine learning tutorial" category "education" author "John Doe"
-
-# Store image with metadata
-LANCE STORE images IMAGE "base64_encoded_image..." filename "photo.jpg" tags "nature,landscape"
-
-# Store multimodal content
-LANCE STORE content TEXT "Product description" IMAGE "base64_image..." type "product"
-```
-
-#### Search Operations
-```bash
-# Search with raw vector
-LANCE SEARCH documents VECTOR "0.1,0.2,0.3,0.4" K 5
-
-# Semantic text search
-LANCE SEARCH.TEXT documents "artificial intelligence" K 10 NPROBES 20
-
-# Generate embeddings
-LANCE EMBED.TEXT "Hello world" "Machine learning"
-```
-
-#### Index Management
-```bash
-# Create performance index
-LANCE CREATE.INDEX documents IVF_PQ PARTITIONS 256 SUBVECTORS 16
-
-# Drop dataset
-LANCE DROP old_dataset
-```
-
-### Configuration
-
-#### Setting Up Embedding Service
-```bash
-# Configure embedding service URL
-redis-cli HSET config:core:aiembed url "http://your-embedding-service:8080/embed"
-
-# Optional: Set authentication token
-redis-cli HSET config:core:aiembed token "your-api-token"
-```
-
-#### Embedding Service API
-Your embedding service should accept POST requests:
-```json
-{
-  "texts": ["text1", "text2"],
-  "images": ["base64_image1", "base64_image2"],
-  "model": "your-model-name"
-}
-```
-
-And return responses:
-```json
-{
-  "embeddings": [[0.1, 0.2, ...], [0.3, 0.4, ...]],
-  "model": "model-name",
-  "usage": {"tokens": 100, "requests": 2}
-}
-```
-
-### Interactive Features
-
-The demo script includes:
-- **Colored output** for better readability
-- **Step-by-step execution** with explanations
-- **Error handling** demonstrations
-- **Automatic cleanup** options
-- **Performance testing** concepts
-- **Real-world usage** examples
-
-### Use Cases Demonstrated
-
-1. **Document Search System**
-   - Semantic document retrieval
-   - Metadata filtering
-   - Relevance ranking
-
-2. **Image Similarity Search**
-   - Visual content matching
-   - Tag-based filtering
-   - Multimodal queries
-
-3. **Product Recommendations**
-   - Feature-based similarity
-   - Category filtering
-   - Price range queries
-
-4. **Content Management**
-   - Mixed media storage
-   - Cross-modal search
-   - Rich metadata support
-
----
+This directory contains examples demonstrating HeroDB's full-text search capabilities powered by Tantivy.
 
 ## Tantivy Search Demo (Bash Script)
 

examples/lance_vector_demo.sh

@@ -1,426 +0,0 @@
-#!/bin/bash
-
-# Lance Vector Database Demo Script
-# This script demonstrates all Lance vector database operations in HeroDB
-
-set -e  # Exit on any error
-
-# Configuration
-REDIS_HOST="localhost"
-REDIS_PORT="6379"
-REDIS_CLI="redis-cli -h $REDIS_HOST -p $REDIS_PORT"
-
-# Colors for output
-RED='\033[0;31m'
-GREEN='\033[0;32m'
-YELLOW='\033[1;33m'
-BLUE='\033[0;34m'
-NC='\033[0m' # No Color
-
-# Helper functions
-log_info() {
-    echo -e "${BLUE}[INFO]${NC} $1"
-}
-
-log_success() {
-    echo -e "${GREEN}[SUCCESS]${NC} $1"
-}
-
-log_warning() {
-    echo -e "${YELLOW}[WARNING]${NC} $1"
-}
-
-log_error() {
-    echo -e "${RED}[ERROR]${NC} $1"
-}
-
-execute_command() {
-    local cmd="$1"
-    local description="$2"
-    
-    echo
-    log_info "Executing: $description"
-    echo "Command: $cmd"
-    
-    if result=$($cmd 2>&1); then
-        log_success "Result: $result"
-    else
-        log_error "Failed: $result"
-        return 1
-    fi
-}
-
-# Check if HeroDB is running
-check_herodb() {
-    log_info "Checking if HeroDB is running..."
-    if ! $REDIS_CLI ping > /dev/null 2>&1; then
-        log_error "HeroDB is not running. Please start it first:"
-        echo "  cargo run -- --dir ./test_data --port $REDIS_PORT"
-        exit 1
-    fi
-    log_success "HeroDB is running"
-}
-
-# Setup embedding service configuration
-setup_embedding_service() {
-    log_info "Setting up embedding service configuration..."
-    
-    # Note: This is a mock URL for demonstration
-    # In production, replace with your actual embedding service
-    execute_command \
-        "$REDIS_CLI HSET config:core:aiembed url 'http://localhost:8080/embed'" \
-        "Configure embedding service URL"
-    
-    # Optional: Set authentication token
-    # execute_command \
-    #     "$REDIS_CLI HSET config:core:aiembed token 'your-api-token'" \
-    #     "Configure embedding service token"
-    
-    log_warning "Note: Embedding service at http://localhost:8080/embed is not running."
-    log_warning "Some operations will fail, but this demonstrates the command structure."
-}
-
-# Dataset Management Operations
-demo_dataset_management() {
-    echo
-    echo "=========================================="
-    echo "         DATASET MANAGEMENT DEMO"
-    echo "=========================================="
-    
-    # List datasets (should be empty initially)
-    execute_command \
-        "$REDIS_CLI LANCE LIST" \
-        "List all datasets (initially empty)"
-    
-    # Create a simple dataset
-    execute_command \
-        "$REDIS_CLI LANCE CREATE documents DIM 384" \
-        "Create a simple document dataset with 384 dimensions"
-    
-    # Create a dataset with schema
-    execute_command \
-        "$REDIS_CLI LANCE CREATE products DIM 768 SCHEMA category:string price:float available:bool description:string" \
-        "Create products dataset with custom schema"
-    
-    # Create an image dataset
-    execute_command \
-        "$REDIS_CLI LANCE CREATE images DIM 512 SCHEMA filename:string tags:string width:int height:int" \
-        "Create images dataset for multimodal content"
-    
-    # List datasets again
-    execute_command \
-        "$REDIS_CLI LANCE LIST" \
-        "List all datasets (should show 3 datasets)"
-    
-    # Get info about datasets
-    execute_command \
-        "$REDIS_CLI LANCE INFO documents" \
-        "Get information about documents dataset"
-    
-    execute_command \
-        "$REDIS_CLI LANCE INFO products" \
-        "Get information about products dataset"
-}
-
-# Embedding Operations
-demo_embedding_operations() {
-    echo
-    echo "=========================================="
-    echo "         EMBEDDING OPERATIONS DEMO"
-    echo "=========================================="
-    
-    log_warning "The following operations will fail because no embedding service is running."
-    log_warning "This demonstrates the command structure and error handling."
-    
-    # Try to embed text (will fail without embedding service)
-    execute_command \
-        "$REDIS_CLI LANCE EMBED.TEXT 'Hello world'" \
-        "Generate embedding for single text" || true
-    
-    # Try to embed multiple texts
-    execute_command \
-        "$REDIS_CLI LANCE EMBED.TEXT 'Machine learning' 'Artificial intelligence' 'Deep learning'" \
-        "Generate embeddings for multiple texts" || true
-}
-
-# Data Storage Operations
-demo_data_storage() {
-    echo
-    echo "=========================================="
-    echo "         DATA STORAGE DEMO"
-    echo "=========================================="
-    
-    log_warning "Storage operations will fail without embedding service, but show command structure."
-    
-    # Store text documents
-    execute_command \
-        "$REDIS_CLI LANCE STORE documents TEXT 'Introduction to machine learning algorithms and their applications in modern AI systems' category 'education' author 'John Doe' difficulty 'beginner'" \
-        "Store a document with text and metadata" || true
-    
-    execute_command \
-        "$REDIS_CLI LANCE STORE documents TEXT 'Deep learning neural networks for computer vision tasks' category 'research' author 'Jane Smith' difficulty 'advanced'" \
-        "Store another document" || true
-    
-    # Store product information
-    execute_command \
-        "$REDIS_CLI LANCE STORE products TEXT 'High-performance laptop with 16GB RAM and SSD storage' category 'electronics' price '1299.99' available 'true'" \
-        "Store product with text description" || true
-    
-    # Store image with metadata (using placeholder base64)
-    execute_command \
-        "$REDIS_CLI LANCE STORE images IMAGE 'iVBORw0KGgoAAAANSUhEUgAAAAEAAAABCAYAAAAfFcSJAAAADUlEQVR42mP8/5+hHgAHggJ/PchI7wAAAABJRU5ErkJggg==' filename 'sample.png' tags 'test,demo' width '1' height '1'" \
-        "Store image with metadata (1x1 pixel PNG)" || true
-    
-    # Store multimodal content
-    execute_command \
-        "$REDIS_CLI LANCE STORE images TEXT 'Beautiful sunset over mountains' IMAGE 'iVBORw0KGgoAAAANSUhEUgAAAAEAAAABCAYAAAAfFcSJAAAADUlEQVR42mP8/5+hHgAHggJ/PchI7wAAAABJRU5ErkJggg==' filename 'sunset.png' tags 'nature,landscape' location 'California'" \
-        "Store multimodal content (text + image)" || true
-}
-
-# Search Operations
-demo_search_operations() {
-    echo
-    echo "=========================================="
-    echo "         SEARCH OPERATIONS DEMO"
-    echo "=========================================="
-    
-    log_warning "Search operations will fail without data, but show command structure."
-    
-    # Search with raw vector
-    execute_command \
-        "$REDIS_CLI LANCE SEARCH documents VECTOR '0.1,0.2,0.3,0.4,0.5' K 5" \
-        "Search with raw vector (5 results)" || true
-    
-    # Search with vector and parameters
-    execute_command \
-        "$REDIS_CLI LANCE SEARCH documents VECTOR '0.1,0.2,0.3,0.4,0.5' K 10 NPROBES 20 REFINE 2" \
-        "Search with vector and advanced parameters" || true
-    
-    # Text-based search
-    execute_command \
-        "$REDIS_CLI LANCE SEARCH.TEXT documents 'machine learning algorithms' K 5" \
-        "Search using text query" || true
-    
-    # Text search with parameters
-    execute_command \
-        "$REDIS_CLI LANCE SEARCH.TEXT products 'laptop computer' K 3 NPROBES 10" \
-        "Search products using text with parameters" || true
-    
-    # Search in image dataset
-    execute_command \
-        "$REDIS_CLI LANCE SEARCH.TEXT images 'sunset landscape' K 5" \
-        "Search images using text description" || true
-}
-
-# Index Management Operations
-demo_index_management() {
-    echo
-    echo "=========================================="
-    echo "         INDEX MANAGEMENT DEMO"
-    echo "=========================================="
-    
-    # Create indexes for better search performance
-    execute_command \
-        "$REDIS_CLI LANCE CREATE.INDEX documents IVF_PQ" \
-        "Create default IVF_PQ index for documents"
-    
-    execute_command \
-        "$REDIS_CLI LANCE CREATE.INDEX products IVF_PQ PARTITIONS 512 SUBVECTORS 32" \
-        "Create IVF_PQ index with custom parameters for products"
-    
-    execute_command \
-        "$REDIS_CLI LANCE CREATE.INDEX images IVF_PQ PARTITIONS 256 SUBVECTORS 16" \
-        "Create IVF_PQ index for images dataset"
-    
-    log_success "Indexes created successfully"
-}
-
-# Advanced Usage Examples
-demo_advanced_usage() {
-    echo
-    echo "=========================================="
-    echo "         ADVANCED USAGE EXAMPLES"
-    echo "=========================================="
-    
-    # Create a specialized dataset for semantic search
-    execute_command \
-        "$REDIS_CLI LANCE CREATE semantic_search DIM 1536 SCHEMA title:string content:string url:string timestamp:string source:string" \
-        "Create dataset for semantic search with rich metadata"
-    
-    # Demonstrate batch operations concept
-    log_info "Batch operations example (would store multiple items):"
-    echo "  for doc in documents:"
-    echo "    LANCE STORE semantic_search TEXT \"\$doc_content\" title \"\$title\" url \"\$url\""
-    
-    # Show monitoring commands
-    log_info "Monitoring and maintenance commands:"
-    execute_command \
-        "$REDIS_CLI LANCE LIST" \
-        "List all datasets for monitoring"
-    
-    # Show dataset statistics
-    for dataset in documents products images semantic_search; do
-        execute_command \
-            "$REDIS_CLI LANCE INFO $dataset" \
-            "Get statistics for $dataset" || true
-    done
-}
-
-# Cleanup Operations
-demo_cleanup() {
-    echo
-    echo "=========================================="
-    echo "         CLEANUP OPERATIONS DEMO"
-    echo "=========================================="
-    
-    log_info "Demonstrating cleanup operations..."
-    
-    # Drop individual datasets
-    execute_command \
-        "$REDIS_CLI LANCE DROP semantic_search" \
-        "Drop semantic_search dataset"
-    
-    # List remaining datasets
-    execute_command \
-        "$REDIS_CLI LANCE LIST" \
-        "List remaining datasets"
-    
-    # Ask user if they want to clean up all test data
-    echo
-    read -p "Do you want to clean up all test datasets? (y/N): " -n 1 -r
-    echo
-    if [[ $REPLY =~ ^[Yy]$ ]]; then
-        execute_command \
-            "$REDIS_CLI LANCE DROP documents" \
-            "Drop documents dataset"
-        
-        execute_command \
-            "$REDIS_CLI LANCE DROP products" \
-            "Drop products dataset"
-        
-        execute_command \
-            "$REDIS_CLI LANCE DROP images" \
-            "Drop images dataset"
-        
-        execute_command \
-            "$REDIS_CLI LANCE LIST" \
-            "Verify all datasets are cleaned up"
-        
-        log_success "All test datasets cleaned up"
-    else
-        log_info "Keeping test datasets for further experimentation"
-    fi
-}
-
-# Error Handling Demo
-demo_error_handling() {
-    echo
-    echo "=========================================="
-    echo "         ERROR HANDLING DEMO"
-    echo "=========================================="
-    
-    log_info "Demonstrating various error conditions..."
-    
-    # Try to access non-existent dataset
-    execute_command \
-        "$REDIS_CLI LANCE INFO nonexistent_dataset" \
-        "Try to get info for non-existent dataset" || true
-    
-    # Try to search non-existent dataset
-    execute_command \
-        "$REDIS_CLI LANCE SEARCH nonexistent_dataset VECTOR '0.1,0.2' K 5" \
-        "Try to search non-existent dataset" || true
-    
-    # Try to drop non-existent dataset
-    execute_command \
-        "$REDIS_CLI LANCE DROP nonexistent_dataset" \
-        "Try to drop non-existent dataset" || true
-    
-    # Try invalid vector format
-    execute_command \
-        "$REDIS_CLI LANCE SEARCH documents VECTOR 'invalid,vector,format' K 5" \
-        "Try search with invalid vector format" || true
-    
-    log_info "Error handling demonstration complete"
-}
-
-# Performance Testing Demo
-demo_performance_testing() {
-    echo
-    echo "=========================================="
-    echo "         PERFORMANCE TESTING DEMO"
-    echo "=========================================="
-    
-    log_info "Creating performance test dataset..."
-    execute_command \
-        "$REDIS_CLI LANCE CREATE perf_test DIM 128 SCHEMA batch_id:string item_id:string" \
-        "Create performance test dataset"
-    
-    log_info "Performance testing would involve:"
-    echo "  1. Bulk loading thousands of vectors"
-    echo "  2. Creating indexes with different parameters"
-    echo "  3. Measuring search latency with various K values"
-    echo "  4. Testing different NPROBES settings"
-    echo "  5. Monitoring memory usage"
-    
-    log_info "Example performance test commands:"
-    echo "  # Test search speed with different parameters"
-    echo "  time redis-cli LANCE SEARCH.TEXT perf_test 'query' K 10"
-    echo "  time redis-cli LANCE SEARCH.TEXT perf_test 'query' K 10 NPROBES 50"
-    echo "  time redis-cli LANCE SEARCH.TEXT perf_test 'query' K 100 NPROBES 100"
-    
-    # Clean up performance test dataset
-    execute_command \
-        "$REDIS_CLI LANCE DROP perf_test" \
-        "Clean up performance test dataset"
-}
-
-# Main execution
-main() {
-    echo "=========================================="
-    echo "    LANCE VECTOR DATABASE DEMO SCRIPT"
-    echo "=========================================="
-    echo
-    echo "This script demonstrates all Lance vector database operations."
-    echo "Note: Some operations will fail without a running embedding service."
-    echo "This is expected and demonstrates error handling."
-    echo
-    
-    # Check prerequisites
-    check_herodb
-    
-    # Setup
-    setup_embedding_service
-    
-    # Run demos
-    demo_dataset_management
-    demo_embedding_operations
-    demo_data_storage
-    demo_search_operations
-    demo_index_management
-    demo_advanced_usage
-    demo_error_handling
-    demo_performance_testing
-    
-    # Cleanup
-    demo_cleanup
-    
-    echo
-    echo "=========================================="
-    echo "         DEMO COMPLETE"
-    echo "=========================================="
-    echo
-    log_success "Lance vector database demo completed successfully!"
-    echo
-    echo "Next steps:"
-    echo "1. Set up a real embedding service (OpenAI, Hugging Face, etc.)"
-    echo "2. Update the embedding service URL configuration"
-    echo "3. Try storing and searching real data"
-    echo "4. Experiment with different vector dimensions and index parameters"
-    echo "5. Build your AI-powered application!"
-    echo
-    echo "For more information, see docs/lance_vector_db.md"
-}
-
-# Run the demo
-main "$@"

src/admin_meta.rs

@@ -0,0 +1,489 @@
+use std::path::PathBuf;
+use std::sync::{Arc, OnceLock, Mutex, RwLock};
+use std::collections::HashMap;
+
+use crate::error::DBError;
+use crate::options;
+use crate::rpc::Permissions;
+use crate::storage::Storage;
+use crate::storage_sled::SledStorage;
+use crate::storage_trait::StorageBackend;
+
+// Key builders
+fn k_admin_next_id() -> &'static str {
+    "admin:next_id"
+}
+fn k_admin_dbs() -> &'static str {
+    "admin:dbs"
+}
+fn k_meta_db(id: u64) -> String {
+    format!("meta:db:{}", id)
+}
+fn k_meta_db_keys(id: u64) -> String {
+    format!("meta:db:{}:keys", id)
+}
+fn k_meta_db_enc(id: u64) -> String {
+    format!("meta:db:{}:enc", id)
+}
+
+// Global cache of admin DB 0 handles per base_dir to avoid sled/reDB file-lock contention
+// and to correctly isolate different test instances with distinct directories.
+static ADMIN_STORAGES: OnceLock<RwLock<HashMap<String, Arc<dyn StorageBackend>>>> = OnceLock::new();
+
+// Global registry for data DB storages to avoid double-open across process.
+static DATA_STORAGES: OnceLock<RwLock<HashMap<u64, Arc<dyn StorageBackend>>>> = OnceLock::new();
+static DATA_INIT_LOCK: Mutex<()> = Mutex::new(());
+
+fn init_admin_storage(
+    base_dir: &str,
+    backend: options::BackendType,
+    admin_secret: &str,
+) -> Result<Arc<dyn StorageBackend>, DBError> {
+    let db_file = PathBuf::from(base_dir).join("0.db");
+    if let Some(parent_dir) = db_file.parent() {
+        std::fs::create_dir_all(parent_dir).map_err(|e| {
+            DBError(format!("Failed to create directory {}: {}", parent_dir.display(), e))
+        })?;
+    }
+    let storage: Arc<dyn StorageBackend> = match backend {
+        options::BackendType::Redb => Arc::new(Storage::new(&db_file, true, Some(admin_secret))?),
+        options::BackendType::Sled => Arc::new(SledStorage::new(&db_file, true, Some(admin_secret))?),
+        options::BackendType::Tantivy => {
+            return Err(DBError("Admin DB 0 cannot use Tantivy backend".to_string()))
+        }
+    };
+    Ok(storage)
+}
+
+// Get or initialize a cached handle to admin DB 0 per base_dir (thread-safe, no double-open race)
+pub fn open_admin_storage(
+    base_dir: &str,
+    backend: options::BackendType,
+    admin_secret: &str,
+) -> Result<Arc<dyn StorageBackend>, DBError> {
+    let map = ADMIN_STORAGES.get_or_init(|| RwLock::new(HashMap::new()));
+    // Fast path
+    if let Some(st) = map.read().unwrap().get(base_dir) {
+        return Ok(st.clone());
+    }
+    // Slow path with write lock
+    {
+        let mut w = map.write().unwrap();
+        if let Some(st) = w.get(base_dir) {
+            return Ok(st.clone());
+        }
+
+        // Detect existing 0.db backend by filesystem, if present.
+        let admin_path = PathBuf::from(base_dir).join("0.db");
+        let detected = if admin_path.exists() {
+            if admin_path.is_file() {
+                Some(options::BackendType::Redb)
+            } else if admin_path.is_dir() {
+                Some(options::BackendType::Sled)
+            } else {
+                None
+            }
+        } else {
+            None
+        };
+
+        let effective_backend = match detected {
+            Some(d) if d != backend => {
+                eprintln!(
+                    "warning: Admin DB 0 at {} appears to be {:?}, but process default is {:?}. Using detected backend.",
+                    admin_path.display(),
+                    d,
+                    backend
+                );
+                d
+            }
+            Some(d) => d,
+            None => backend, // First boot: use requested backend to initialize 0.db
+        };
+
+        let st = init_admin_storage(base_dir, effective_backend, admin_secret)?;
+        w.insert(base_dir.to_string(), st.clone());
+        Ok(st)
+    }
+}
+
+// Ensure admin structures exist in encrypted DB 0
+pub fn ensure_bootstrap(
+    base_dir: &str,
+    backend: options::BackendType,
+    admin_secret: &str,
+) -> Result<(), DBError> {
+    let admin = open_admin_storage(base_dir, backend, admin_secret)?;
+
+    // Initialize next id if missing
+    if !admin.exists(k_admin_next_id())? {
+        admin.set(k_admin_next_id().to_string(), "1".to_string())?;
+    }
+    // admin:dbs is a hash; it's fine if it doesn't exist (hlen -> 0)
+    Ok(())
+}
+
+// Get or initialize a shared handle to a data DB (> 0), avoiding double-open across subsystems
+pub fn open_data_storage(
+    base_dir: &str,
+    backend: options::BackendType,
+    admin_secret: &str,
+    id: u64,
+) -> Result<Arc<dyn StorageBackend>, DBError> {
+    if id == 0 {
+        return open_admin_storage(base_dir, backend, admin_secret);
+    }
+
+    // Validate existence in admin metadata
+    if !db_exists(base_dir, backend.clone(), admin_secret, id)? {
+        return Err(DBError(format!(
+            "Cannot open database instance {}, as that database instance does not exist.",
+            id
+        )));
+    }
+
+    let map = DATA_STORAGES.get_or_init(|| RwLock::new(HashMap::new()));
+    // Fast path
+    if let Some(st) = map.read().unwrap().get(&id) {
+        return Ok(st.clone());
+    }
+
+    // Slow path with init lock
+    let _guard = DATA_INIT_LOCK.lock().unwrap();
+    if let Some(st) = map.read().unwrap().get(&id) {
+        return Ok(st.clone());
+    }
+
+    // Resolve effective backend for this db id:
+    // 1) Try admin meta "backend" field
+    // 2) If missing, sniff filesystem (file => Redb, dir => Sled), then persist into admin meta
+    // 3) Fallback to requested 'backend' (startup default) if nothing else is known
+    let meta_backend = get_database_backend(base_dir, backend.clone(), admin_secret, id).ok().flatten();
+    let db_path = PathBuf::from(base_dir).join(format!("{}.db", id));
+    let sniffed_backend = if db_path.exists() {
+        if db_path.is_file() {
+            Some(options::BackendType::Redb)
+        } else if db_path.is_dir() {
+            Some(options::BackendType::Sled)
+        } else {
+            None
+        }
+    } else {
+        None
+    };
+    let effective_backend = meta_backend.clone().or(sniffed_backend).unwrap_or(backend.clone());
+
+    // If we had to sniff (i.e., meta missing), persist it for future robustness
+    if meta_backend.is_none() {
+        let _ = set_database_backend(base_dir, backend.clone(), admin_secret, id, effective_backend.clone());
+    }
+
+    // Warn if caller-provided backend differs from effective
+    if effective_backend != backend {
+        eprintln!(
+            "notice: Database {} backend resolved to {:?} (caller requested {:?}). Using resolved backend.",
+            id, effective_backend, backend
+        );
+    }
+
+    // Determine per-db encryption (from admin meta)
+    let enc = get_enc_key(base_dir, backend.clone(), admin_secret, id)?;
+    let should_encrypt = enc.is_some();
+
+    // Build database file path and ensure parent dir exists
+    let db_file = PathBuf::from(base_dir).join(format!("{}.db", id));
+    if let Some(parent_dir) = db_file.parent() {
+        std::fs::create_dir_all(parent_dir).map_err(|e| {
+            DBError(format!("Failed to create directory {}: {}", parent_dir.display(), e))
+        })?;
+    }
+
+    // Open storage using the effective backend
+    let storage: Arc<dyn StorageBackend> = match effective_backend {
+        options::BackendType::Redb => Arc::new(Storage::new(&db_file, should_encrypt, enc.as_deref())?),
+        options::BackendType::Sled => Arc::new(SledStorage::new(&db_file, should_encrypt, enc.as_deref())?),
+        options::BackendType::Tantivy => {
+            return Err(DBError("Tantivy backend has no KV storage; use FT.* commands only".to_string()))
+        }
+    };
+
+    // Publish to registry
+    map.write().unwrap().insert(id, storage.clone());
+    Ok(storage)
+}
+
+// Allocate the next DB id and persist new pointer
+pub fn allocate_next_id(
+    base_dir: &str,
+    backend: options::BackendType,
+    admin_secret: &str,
+) -> Result<u64, DBError> {
+    let admin = open_admin_storage(base_dir, backend, admin_secret)?;
+    let cur = admin
+        .get(k_admin_next_id())?
+        .unwrap_or_else(|| "1".to_string());
+    let id: u64 = cur.parse().unwrap_or(1);
+    let next = id.checked_add(1).ok_or_else(|| DBError("next_id overflow".into()))?;
+    admin.set(k_admin_next_id().to_string(), next.to_string())?;
+
+    // Register into admin:dbs set/hash
+    let _ = admin.hset(k_admin_dbs(), vec![(id.to_string(), "1".to_string())])?;
+
+    // Default meta for the new db: public true
+    let meta_key = k_meta_db(id);
+    let _ = admin.hset(&meta_key, vec![("public".to_string(), "true".to_string())])?;
+
+    Ok(id)
+}
+
+// Check existence of a db id in admin:dbs
+pub fn db_exists(
+    base_dir: &str,
+    backend: options::BackendType,
+    admin_secret: &str,
+    id: u64,
+) -> Result<bool, DBError> {
+    let admin = open_admin_storage(base_dir, backend, admin_secret)?;
+    Ok(admin.hexists(k_admin_dbs(), &id.to_string())?)
+}
+
+// Get per-db encryption key, if any
+pub fn get_enc_key(
+    base_dir: &str,
+    backend: options::BackendType,
+    admin_secret: &str,
+    id: u64,
+) -> Result<Option<String>, DBError> {
+    let admin = open_admin_storage(base_dir, backend, admin_secret)?;
+    admin.get(&k_meta_db_enc(id))
+}
+
+// Set per-db encryption key (called during create)
+pub fn set_enc_key(
+    base_dir: &str,
+    backend: options::BackendType,
+    admin_secret: &str,
+    id: u64,
+    key: &str,
+) -> Result<(), DBError> {
+    let admin = open_admin_storage(base_dir, backend, admin_secret)?;
+    admin.set(k_meta_db_enc(id), key.to_string())
+}
+
+// Set database public flag
+pub fn set_database_public(
+    base_dir: &str,
+    backend: options::BackendType,
+    admin_secret: &str,
+    id: u64,
+    public: bool,
+) -> Result<(), DBError> {
+    let admin = open_admin_storage(base_dir, backend, admin_secret)?;
+    let mk = k_meta_db(id);
+    let _ = admin.hset(&mk, vec![("public".to_string(), public.to_string())])?;
+    Ok(())
+}
+
+// Persist per-db backend type in admin metadata (module-scope)
+pub fn set_database_backend(
+    base_dir: &str,
+    backend: options::BackendType,
+    admin_secret: &str,
+    id: u64,
+    db_backend: options::BackendType,
+) -> Result<(), DBError> {
+    let admin = open_admin_storage(base_dir, backend, admin_secret)?;
+    let mk = k_meta_db(id);
+    let val = match db_backend {
+        options::BackendType::Redb => "Redb",
+        options::BackendType::Sled => "Sled",
+        options::BackendType::Tantivy => "Tantivy",
+    };
+    let _ = admin.hset(&mk, vec![("backend".to_string(), val.to_string())])?;
+    Ok(())
+}
+
+pub fn get_database_backend(
+    base_dir: &str,
+    backend: options::BackendType,
+    admin_secret: &str,
+    id: u64,
+) -> Result<Option<options::BackendType>, DBError> {
+    let admin = open_admin_storage(base_dir, backend, admin_secret)?;
+    let mk = k_meta_db(id);
+    match admin.hget(&mk, "backend")? {
+        Some(s) if s == "Redb" => Ok(Some(options::BackendType::Redb)),
+        Some(s) if s == "Sled" => Ok(Some(options::BackendType::Sled)),
+        Some(s) if s == "Tantivy" => Ok(Some(options::BackendType::Tantivy)),
+        _ => Ok(None),
+    }
+}
+
+// Set database name
+pub fn set_database_name(
+    base_dir: &str,
+    backend: options::BackendType,
+    admin_secret: &str,
+    id: u64,
+    name: &str,
+) -> Result<(), DBError> {
+    let admin = open_admin_storage(base_dir, backend, admin_secret)?;
+    let mk = k_meta_db(id);
+    let _ = admin.hset(&mk, vec![("name".to_string(), name.to_string())])?;
+    Ok(())
+}
+
+// Get database name
+pub fn get_database_name(
+    base_dir: &str,
+    backend: options::BackendType,
+    admin_secret: &str,
+    id: u64,
+) -> Result<Option<String>, DBError> {
+    let admin = open_admin_storage(base_dir, backend, admin_secret)?;
+    let mk = k_meta_db(id);
+    admin.hget(&mk, "name")
+}
+
+// Internal: load public flag; default to true when meta missing
+fn load_public(
+    admin: &Arc<dyn StorageBackend>,
+    id: u64,
+) -> Result<bool, DBError> {
+    let mk = k_meta_db(id);
+    match admin.hget(&mk, "public")? {
+        Some(v) => Ok(v == "true"),
+        None => Ok(true),
+    }
+}
+
+// Add access key for db (value format: "Read:ts" or "ReadWrite:ts")
+pub fn add_access_key(
+    base_dir: &str,
+    backend: options::BackendType,
+    admin_secret: &str,
+    id: u64,
+    key_plain: &str,
+    perms: Permissions,
+) -> Result<(), DBError> {
+    let admin = open_admin_storage(base_dir, backend, admin_secret)?;
+    let hash = crate::rpc::hash_key(key_plain);
+    let v = match perms {
+        Permissions::Read => format!("Read:{}", now_secs()),
+        Permissions::ReadWrite => format!("ReadWrite:{}", now_secs()),
+    };
+    let _ = admin.hset(&k_meta_db_keys(id), vec![(hash, v)])?;
+    Ok(())
+}
+
+// Delete access key by hash
+pub fn delete_access_key(
+    base_dir: &str,
+    backend: options::BackendType,
+    admin_secret: &str,
+    id: u64,
+    key_hash: &str,
+) -> Result<bool, DBError> {
+    let admin = open_admin_storage(base_dir, backend, admin_secret)?;
+    let n = admin.hdel(&k_meta_db_keys(id), vec![key_hash.to_string()])?;
+    Ok(n > 0)
+}
+
+// List access keys, returning (hash, perms, created_at_secs)
+pub fn list_access_keys(
+    base_dir: &str,
+    backend: options::BackendType,
+    admin_secret: &str,
+    id: u64,
+) -> Result<Vec<(String, Permissions, u64)>, DBError> {
+    let admin = open_admin_storage(base_dir, backend, admin_secret)?;
+    let pairs = admin.hgetall(&k_meta_db_keys(id))?;
+    let mut out = Vec::new();
+    for (hash, val) in pairs {
+        let (perm, ts) = parse_perm_value(&val);
+        out.push((hash, perm, ts));
+    }
+    Ok(out)
+}
+
+// Verify access permission for db id with optional key
+// Returns:
+// - Ok(Some(Permissions)) when access is allowed
+// - Ok(None) when not allowed or db missing (caller can distinguish by calling db_exists)
+pub fn verify_access(
+    base_dir: &str,
+    backend: options::BackendType,
+    admin_secret: &str,
+    id: u64,
+    key_opt: Option<&str>,
+) -> Result<Option<Permissions>, DBError> {
+    // Admin DB 0: require exact admin_secret
+    if id == 0 {
+        if let Some(k) = key_opt {
+            if k == admin_secret {
+                return Ok(Some(Permissions::ReadWrite));
+            }
+        }
+        return Ok(None);
+    }
+
+    let admin = open_admin_storage(base_dir, backend, admin_secret)?;
+    if !admin.hexists(k_admin_dbs(), &id.to_string())? {
+        return Ok(None);
+    }
+
+    // Public?
+    if load_public(&admin, id)? {
+        return Ok(Some(Permissions::ReadWrite));
+    }
+
+    // Private: require key and verify
+    if let Some(k) = key_opt {
+        let hash = crate::rpc::hash_key(k);
+        if let Some(v) = admin.hget(&k_meta_db_keys(id), &hash)? {
+            let (perm, _ts) = parse_perm_value(&v);
+            return Ok(Some(perm));
+        }
+    }
+    Ok(None)
+}
+
+// Enumerate all db ids
+pub fn list_dbs(
+    base_dir: &str,
+    backend: options::BackendType,
+    admin_secret: &str,
+) -> Result<Vec<u64>, DBError> {
+    let admin = open_admin_storage(base_dir, backend, admin_secret)?;
+    let ids = admin.hkeys(k_admin_dbs())?;
+    let mut out = Vec::new();
+    for s in ids {
+        if let Ok(v) = s.parse() {
+            out.push(v);
+        }
+    }
+    Ok(out)
+}
+
+// Helper: parse permission value "Read:ts" or "ReadWrite:ts"
+fn parse_perm_value(v: &str) -> (Permissions, u64) {
+    let mut parts = v.split(':');
+    let p = parts.next().unwrap_or("Read");
+    let ts = parts
+        .next()
+        .and_then(|s| s.parse().ok())
+        .unwrap_or(0u64);
+    let perm = match p {
+        "ReadWrite" => Permissions::ReadWrite,
+        _ => Permissions::Read,
+    };
+    (perm, ts)
+}
+
+fn now_secs() -> u64 {
+    use std::time::{SystemTime, UNIX_EPOCH};
+    SystemTime::now()
+        .duration_since(UNIX_EPOCH)
+        .unwrap_or_default()
+        .as_secs()
+}

src/age.rs

@@ -19,6 +19,8 @@ use age::x25519;
 use ed25519_dalek::{Signature, Signer, Verifier, SigningKey, VerifyingKey};
 
 use base64::{engine::general_purpose::STANDARD as B64, Engine as _};
+use std::collections::HashSet;
+use std::convert::TryInto;
 
 use crate::protocol::Protocol;
 use crate::server::Server;
@@ -74,6 +76,125 @@ fn parse_ed25519_verifying_key(s: &str) -> Result<VerifyingKey, AgeWireError> {
     VerifyingKey::from_bytes(&key_bytes).map_err(|_| AgeWireError::ParseKey)
 }
 
+// ---------- Derivation + Raw X25519 (Ed25519 -> X25519) ----------
+//
+// We deterministically derive an X25519 keypair from an Ed25519 SigningKey.
+// We persist the X25519 public/secret as base64-encoded 32-byte raw values
+// (no "age1..."/"AGE-SECRET-KEY-1..." formatting). Name-based encrypt/decrypt
+// uses these raw values directly via x25519-dalek + ChaCha20Poly1305.
+
+use chacha20poly1305::{aead::{Aead, KeyInit}, ChaCha20Poly1305, Key, Nonce};
+use sha2::{Digest, Sha256};
+use x25519_dalek::{PublicKey as XPublicKey, StaticSecret as XStaticSecret};
+
+fn derive_x25519_raw_from_ed25519(sk: &SigningKey) -> ([u8; 32], [u8; 32]) {
+    // X25519 secret scalar (clamped) from Ed25519 secret
+    let scalar: [u8; 32] = sk.to_scalar_bytes();
+    // Build X25519 secret/public using dalek
+    let xsec = XStaticSecret::from(scalar);
+    let xpub = XPublicKey::from(&xsec);
+    (xpub.to_bytes(), xsec.to_bytes())
+}
+
+fn derive_x25519_raw_b64_from_ed25519(sk: &SigningKey) -> (String, String) {
+    let (xpub, xsec) = derive_x25519_raw_from_ed25519(sk);
+    (B64.encode(xpub), B64.encode(xsec))
+}
+
+// Helper: detect whether a stored key looks like an age-formatted string
+fn looks_like_age_format(s: &str) -> bool {
+    s.starts_with("age1") || s.starts_with("AGE-SECRET-KEY-1")
+}
+
+// Our container format for name-based raw X25519 encryption:
+// bytes = "HDBX1" (5) || eph_pub(32) || nonce(12) || ciphertext(..)
+// Entire blob is base64-encoded for transport.
+const HDBX1_MAGIC: &[u8; 5] = b"HDBX1";
+
+fn encrypt_b64_with_x25519_raw(recip_pub_b64: &str, msg: &str) -> Result<String, AgeWireError> {
+    use rand::RngCore;
+    use rand::rngs::OsRng;
+
+    // Parse recipient public key (raw 32 bytes, base64)
+    let recip_pub_bytes = B64.decode(recip_pub_b64).map_err(|_| AgeWireError::ParseKey)?;
+    if recip_pub_bytes.len() != 32 { return Err(AgeWireError::ParseKey); }
+    let recip_pub_arr: [u8; 32] = recip_pub_bytes.as_slice().try_into().map_err(|_| AgeWireError::ParseKey)?;
+    let recip_pub: XPublicKey = XPublicKey::from(recip_pub_arr);
+
+    // Generate ephemeral X25519 keypair
+    let mut eph_sec_bytes = [0u8; 32];
+    OsRng.fill_bytes(&mut eph_sec_bytes);
+    let eph_sec = XStaticSecret::from(eph_sec_bytes);
+    let eph_pub = XPublicKey::from(&eph_sec);
+
+    // ECDH
+    let shared = eph_sec.diffie_hellman(&recip_pub);
+    // Derive symmetric key via SHA-256 over context + shared + parties
+    let mut hasher = Sha256::default();
+    hasher.update(b"herodb-x25519-v1");
+    hasher.update(shared.as_bytes());
+    hasher.update(eph_pub.as_bytes());
+    hasher.update(recip_pub.as_bytes());
+    let key_bytes = hasher.finalize();
+    let key = Key::from_slice(&key_bytes[..32]);
+
+    // Nonce (12 bytes)
+    let mut nonce_bytes = [0u8; 12];
+    OsRng.fill_bytes(&mut nonce_bytes);
+    let nonce = Nonce::from_slice(&nonce_bytes);
+
+    // Encrypt
+    let cipher = ChaCha20Poly1305::new(key);
+    let ct = cipher.encrypt(nonce, msg.as_bytes())
+        .map_err(|e| AgeWireError::Crypto(format!("encrypt: {e}")))?;
+
+    // Assemble container
+    let mut out = Vec::with_capacity(5 + 32 + 12 + ct.len());
+    out.extend_from_slice(HDBX1_MAGIC);
+    out.extend_from_slice(eph_pub.as_bytes());
+    out.extend_from_slice(&nonce_bytes);
+    out.extend_from_slice(&ct);
+
+    Ok(B64.encode(out))
+}
+
+fn decrypt_b64_with_x25519_raw(identity_sec_b64: &str, ct_b64: &str) -> Result<String, AgeWireError> {
+    // Parse X25519 secret (raw 32 bytes, base64)
+    let sec_bytes = B64.decode(identity_sec_b64).map_err(|_| AgeWireError::ParseKey)?;
+    if sec_bytes.len() != 32 { return Err(AgeWireError::ParseKey); }
+    let sec_arr: [u8; 32] = sec_bytes.as_slice().try_into().map_err(|_| AgeWireError::ParseKey)?;
+    let xsec = XStaticSecret::from(sec_arr);
+    let xpub = XPublicKey::from(&xsec); // self public
+
+    // Decode container
+    let blob = B64.decode(ct_b64.as_bytes()).map_err(|e| AgeWireError::Crypto(e.to_string()))?;
+    if blob.len() < 5 + 32 + 12 { return Err(AgeWireError::Crypto("ciphertext too short".to_string())); }
+    if &blob[..5] != HDBX1_MAGIC { return Err(AgeWireError::Crypto("bad header".to_string())); }
+
+    let eph_pub_arr: [u8; 32] = blob[5..5+32].try_into().map_err(|_| AgeWireError::Crypto("bad eph pub".to_string()))?;
+    let eph_pub = XPublicKey::from(eph_pub_arr);
+    let nonce_bytes: [u8; 12] = blob[5+32..5+32+12].try_into().unwrap();
+    let ct = &blob[5+32+12..];
+
+    // Recompute shared + key
+    let shared = xsec.diffie_hellman(&eph_pub);
+    let mut hasher = Sha256::default();
+    hasher.update(b"herodb-x25519-v1");
+    hasher.update(shared.as_bytes());
+    hasher.update(eph_pub.as_bytes());
+    hasher.update(xpub.as_bytes());
+    let key_bytes = hasher.finalize();
+    let key = Key::from_slice(&key_bytes[..32]);
+
+    // Decrypt
+    let cipher = ChaCha20Poly1305::new(key);
+    let nonce = Nonce::from_slice(&nonce_bytes);
+    let pt = cipher.decrypt(nonce, ct)
+        .map_err(|e| AgeWireError::Crypto(format!("decrypt: {e}")))?;
+
+    String::from_utf8(pt).map_err(|_| AgeWireError::Utf8)
+}
+
 // ---------- Stateless crypto helpers (string in/out) ----------
 
 pub fn gen_enc_keypair() -> (String, String) {
@@ -210,13 +331,72 @@ pub async fn cmd_age_verify(verify_pub: &str, message: &str, sig_b64: &str) -> P
     }
 }
 
+// ---------- NEW: unified stateless generator (Ed25519 + derived X25519 raw) ----------
+//
+// Returns 4-tuple:
+// [ verify_pub_b64 (32B), signpriv_b64 (32B), x25519_pub_b64 (32B), x25519_sec_b64 (32B) ]
+// No persistence (stateless).
+pub async fn cmd_age_genkey() -> Protocol {
+    use rand::RngCore;
+    use rand::rngs::OsRng;
+
+    let mut secret_bytes = [0u8; 32];
+    OsRng.fill_bytes(&mut secret_bytes);
+
+    let signing_key = SigningKey::from_bytes(&secret_bytes);
+    let verifying_key = signing_key.verifying_key();
+
+    let verify_b64 = B64.encode(verifying_key.to_bytes());
+    let sign_b64 = B64.encode(signing_key.to_bytes());
+
+    let (xpub_b64, xsec_b64) = derive_x25519_raw_b64_from_ed25519(&signing_key);
+
+    Protocol::Array(vec![
+        Protocol::BulkString(verify_b64),
+        Protocol::BulkString(sign_b64),
+        Protocol::BulkString(xpub_b64),
+        Protocol::BulkString(xsec_b64),
+    ])
+}
+
 // ---------- NEW: Persistent, named-key commands ----------
 
 pub async fn cmd_age_keygen(server: &Server, name: &str) -> Protocol {
-    let (recip, ident) = gen_enc_keypair();
-    if let Err(e) = sset(server, &enc_pub_key_key(name), &recip) { return e.to_protocol(); }
-    if let Err(e) = sset(server, &enc_priv_key_key(name), &ident) { return e.to_protocol(); }
-    Protocol::Array(vec![Protocol::BulkString(recip), Protocol::BulkString(ident)])
+    use rand::RngCore;
+    use rand::rngs::OsRng;
+
+    // Generate Ed25519 keypair
+    let mut secret_bytes = [0u8; 32];
+    OsRng.fill_bytes(&mut secret_bytes);
+    let signing_key = SigningKey::from_bytes(&secret_bytes);
+    let verifying_key = signing_key.verifying_key();
+
+    // Encode Ed25519 as base64 (32 bytes)
+    let verify_b64 = B64.encode(verifying_key.to_bytes());
+    let sign_b64 = B64.encode(signing_key.to_bytes());
+
+    // Derive X25519 raw (32-byte) keys and encode as base64
+    let (xpub_b64, xsec_b64) = derive_x25519_raw_b64_from_ed25519(&signing_key);
+
+    // Decode to create age-formatted strings
+    let xpub_bytes = B64.decode(&xpub_b64).unwrap();
+    let xsec_bytes = B64.decode(&xsec_b64).unwrap();
+    let xpub_arr: [u8; 32] = xpub_bytes.as_slice().try_into().unwrap();
+    let xsec_arr: [u8; 32] = xsec_bytes.as_slice().try_into().unwrap();
+    let recip_str = format!("age1{}", B64.encode(xpub_arr));
+    let ident_str = format!("AGE-SECRET-KEY-1{}", B64.encode(xsec_arr));
+
+    // Persist Ed25519 and derived X25519 (key-managed mode)
+    if let Err(e) = sset(server, &sign_pub_key_key(name), &verify_b64) { return e.to_protocol(); }
+    if let Err(e) = sset(server, &sign_priv_key_key(name), &sign_b64) { return e.to_protocol(); }
+    if let Err(e) = sset(server, &enc_pub_key_key(name), &xpub_b64) { return e.to_protocol(); }
+    if let Err(e) = sset(server, &enc_priv_key_key(name), &xsec_b64) { return e.to_protocol(); }
+
+    // Return [recipient, identity] in age format
+    Protocol::Array(vec![
+        Protocol::BulkString(recip_str),
+        Protocol::BulkString(ident_str),
+    ])
 }
 
 pub async fn cmd_age_signkeygen(server: &Server, name: &str) -> Protocol {
@@ -227,26 +407,76 @@ pub async fn cmd_age_signkeygen(server: &Server, name: &str) -> Protocol {
 }
 
 pub async fn cmd_age_encrypt_name(server: &Server, name: &str, message: &str) -> Protocol {
-    let recip = match sget(server, &enc_pub_key_key(name)) {
+    // Load stored recipient (could be raw b64 32-byte or "age1..." from legacy)
+    let recip_or_b64 = match sget(server, &enc_pub_key_key(name)) {
         Ok(Some(v)) => v,
-        Ok(None) => return AgeWireError::NotFound("recipient (age:key:{name})").to_protocol(),
+        Ok(None) => {
+            // Derive from stored Ed25519 if present, then persist
+            match sget(server, &sign_priv_key_key(name)) {
+                Ok(Some(sign_b64)) => {
+                    let sk = match parse_ed25519_signing_key(&sign_b64) {
+                        Ok(k) => k,
+                        Err(e) => return e.to_protocol(),
+                    };
+                    let (xpub_b64, xsec_b64) = derive_x25519_raw_b64_from_ed25519(&sk);
+                    if let Err(e) = sset(server, &enc_pub_key_key(name), &xpub_b64) { return e.to_protocol(); }
+                    if let Err(e) = sset(server, &enc_priv_key_key(name), &xsec_b64) { return e.to_protocol(); }
+                    xpub_b64
+                }
+                Ok(None) => return AgeWireError::NotFound("recipient (age:key:{name})").to_protocol(),
+                Err(e) => return e.to_protocol(),
+            }
+        }
         Err(e) => return e.to_protocol(),
     };
-    match encrypt_b64(&recip, message) {
-        Ok(ct) => Protocol::BulkString(ct),
-        Err(e) => e.to_protocol(),
+
+    if looks_like_age_format(&recip_or_b64) {
+        match encrypt_b64(&recip_or_b64, message) {
+            Ok(ct) => Protocol::BulkString(ct),
+            Err(e) => e.to_protocol(),
+        }
+    } else {
+        match encrypt_b64_with_x25519_raw(&recip_or_b64, message) {
+            Ok(ct) => Protocol::BulkString(ct),
+            Err(e) => e.to_protocol(),
+        }
     }
 }
 
 pub async fn cmd_age_decrypt_name(server: &Server, name: &str, ct_b64: &str) -> Protocol {
-    let ident = match sget(server, &enc_priv_key_key(name)) {
+    // Load stored identity (could be raw b64 32-byte or "AGE-SECRET-KEY-1..." from legacy)
+    let ident_or_b64 = match sget(server, &enc_priv_key_key(name)) {
         Ok(Some(v)) => v,
-        Ok(None) => return AgeWireError::NotFound("identity (age:privkey:{name})").to_protocol(),
+        Ok(None) => {
+            // Derive from stored Ed25519 if present, then persist
+            match sget(server, &sign_priv_key_key(name)) {
+                Ok(Some(sign_b64)) => {
+                    let sk = match parse_ed25519_signing_key(&sign_b64) {
+                        Ok(k) => k,
+                        Err(e) => return e.to_protocol(),
+                    };
+                    let (xpub_b64, xsec_b64) = derive_x25519_raw_b64_from_ed25519(&sk);
+                    if let Err(e) = sset(server, &enc_pub_key_key(name), &xpub_b64) { return e.to_protocol(); }
+                    if let Err(e) = sset(server, &enc_priv_key_key(name), &xsec_b64) { return e.to_protocol(); }
+                    xsec_b64
+                }
+                Ok(None) => return AgeWireError::NotFound("identity (age:privkey:{name})").to_protocol(),
+                Err(e) => return e.to_protocol(),
+            }
+        }
         Err(e) => return e.to_protocol(),
     };
-    match decrypt_b64(&ident, ct_b64) {
-        Ok(pt) => Protocol::BulkString(pt),
-        Err(e) => e.to_protocol(),
+
+    if looks_like_age_format(&ident_or_b64) {
+        match decrypt_b64(&ident_or_b64, ct_b64) {
+            Ok(pt) => Protocol::BulkString(pt),
+            Err(e) => e.to_protocol(),
+        }
+    } else {
+        match decrypt_b64_with_x25519_raw(&ident_or_b64, ct_b64) {
+            Ok(pt) => Protocol::BulkString(pt),
+            Err(e) => e.to_protocol(),
+        }
     }
 }
 
@@ -276,33 +506,31 @@ pub async fn cmd_age_verify_name(server: &Server, name: &str, message: &str, sig
 }
 
 pub async fn cmd_age_list(server: &Server) -> Protocol {
-    // Returns 4 arrays: ["encpub", <names...>], ["encpriv", ...], ["signpub", ...], ["signpriv", ...]
+    // Return a flat, deduplicated, sorted list of managed key names (no labels)
     let st = match server.current_storage() { Ok(s) => s, Err(e) => return Protocol::err(&e.0) };
 
     let pull = |pat: &str, prefix: &str| -> Result<Vec<String>, DBError> {
         let keys = st.keys(pat)?;
-        let mut names: Vec<String> = keys.into_iter()
+        let mut names: Vec<String> = keys
+            .into_iter()
             .filter_map(|k| k.strip_prefix(prefix).map(|x| x.to_string()))
             .collect();
         names.sort();
         Ok(names)
     };
 
-    let encpub  = match pull("age:key:*",      "age:key:")      { Ok(v) => v, Err(e)=> return Protocol::err(&e.0) };
-    let encpriv = match pull("age:privkey:*",  "age:privkey:")  { Ok(v) => v, Err(e)=> return Protocol::err(&e.0) };
-    let signpub = match pull("age:signpub:*",  "age:signpub:")  { Ok(v) => v, Err(e)=> return Protocol::err(&e.0) };
-    let signpriv= match pull("age:signpriv:*", "age:signpriv:") { Ok(v) => v, Err(e)=> return Protocol::err(&e.0) };
+    let encpub   = match pull("age:key:*",      "age:key:")      { Ok(v) => v, Err(e)=> return Protocol::err(&e.0) };
+    let encpriv  = match pull("age:privkey:*",  "age:privkey:")  { Ok(v) => v, Err(e)=> return Protocol::err(&e.0) };
+    let signpub  = match pull("age:signpub:*",  "age:signpub:")  { Ok(v) => v, Err(e)=> return Protocol::err(&e.0) };
+    let signpriv = match pull("age:signpriv:*", "age:signpriv:") { Ok(v) => v, Err(e)=> return Protocol::err(&e.0) };
 
-    let to_arr = |label: &str, v: Vec<String>| {
-        let mut out = vec![Protocol::BulkString(label.to_string())];
-        out.push(Protocol::Array(v.into_iter().map(Protocol::BulkString).collect()));
-        Protocol::Array(out)
-    };
+    let mut set: HashSet<String> = HashSet::new();
+    for n in encpub.into_iter().chain(encpriv).chain(signpub).chain(signpriv) {
+        set.insert(n);
+    }
 
-    Protocol::Array(vec![
-        to_arr("encpub", encpub),
-        to_arr("encpriv", encpriv),
-        to_arr("signpub", signpub),
-        to_arr("signpriv", signpriv),
-    ])
+    let mut names: Vec<String> = set.into_iter().collect();
+    names.sort();
+
+    Protocol::Array(names.into_iter().map(Protocol::BulkString).collect())
 }

src/error.rs

@@ -9,12 +9,6 @@ use bincode;
 #[derive(Debug)]
 pub struct DBError(pub String);
 
-impl std::fmt::Display for DBError {
-    fn fmt(&self, f: &mut std::fmt::Formatter<'_>) -> std::fmt::Result {
-        write!(f, "{}", self.0)
-    }
-}
-
 impl From<std::io::Error> for DBError {
     fn from(item: std::io::Error) -> Self {
         DBError(item.to_string().clone())
@@ -98,40 +92,3 @@ impl From<chacha20poly1305::Error> for DBError {
         DBError(item.to_string())
     }
 }
-
-// Lance and related dependencies error handling
-impl From<lance::Error> for DBError {
-    fn from(item: lance::Error) -> Self {
-        DBError(item.to_string())
-    }
-}
-
-impl From<arrow::error::ArrowError> for DBError {
-    fn from(item: arrow::error::ArrowError) -> Self {
-        DBError(item.to_string())
-    }
-}
-
-impl From<reqwest::Error> for DBError {
-    fn from(item: reqwest::Error) -> Self {
-        DBError(item.to_string())
-    }
-}
-
-impl From<image::ImageError> for DBError {
-    fn from(item: image::ImageError) -> Self {
-        DBError(item.to_string())
-    }
-}
-
-impl From<uuid::Error> for DBError {
-    fn from(item: uuid::Error) -> Self {
-        DBError(item.to_string())
-    }
-}
-
-impl From<base64::DecodeError> for DBError {
-    fn from(item: base64::DecodeError) -> Self {
-        DBError(item.to_string())
-    }
-}

src/lance_store.rs

@@ -1,609 +0,0 @@
-use std::collections::HashMap;
-use std::path::PathBuf;
-use std::sync::Arc;
-use tokio::sync::RwLock;
-
-use arrow::array::{Float32Array, StringArray, ArrayRef, FixedSizeListArray, Array};
-use arrow::datatypes::{DataType, Field, Schema, SchemaRef};
-use arrow::record_batch::{RecordBatch, RecordBatchReader};
-use arrow::error::ArrowError;
-use lance::dataset::{Dataset, WriteParams, WriteMode};
-use lance::index::vector::VectorIndexParams;
-use lance_index::vector::pq::PQBuildParams;
-use lance_index::vector::ivf::IvfBuildParams;
-use lance_index::DatasetIndexExt;
-use lance_linalg::distance::MetricType;
-use futures::TryStreamExt;
-use base64::Engine;
-
-use serde::{Deserialize, Serialize};
-use crate::error::DBError;
-
-// Simple RecordBatchReader implementation for Vec<RecordBatch>
-struct VecRecordBatchReader {
-    batches: std::vec::IntoIter<Result<RecordBatch, ArrowError>>,
-}
-
-impl VecRecordBatchReader {
-    fn new(batches: Vec<RecordBatch>) -> Self {
-        let result_batches = batches.into_iter().map(Ok).collect::<Vec<_>>();
-        Self {
-            batches: result_batches.into_iter(),
-        }
-    }
-}
-
-impl Iterator for VecRecordBatchReader {
-    type Item = Result<RecordBatch, ArrowError>;
-
-    fn next(&mut self) -> Option<Self::Item> {
-        self.batches.next()
-    }
-}
-
-impl RecordBatchReader for VecRecordBatchReader {
-    fn schema(&self) -> SchemaRef {
-        // This is a simplified implementation - in practice you'd want to store the schema
-        Arc::new(Schema::empty())
-    }
-}
-
-#[derive(Debug, Serialize, Deserialize)]
-struct EmbeddingRequest {
-    texts: Option<Vec<String>>,
-    images: Option<Vec<String>>, // base64 encoded
-    model: Option<String>,
-}
-
-#[derive(Debug, Serialize, Deserialize)]
-struct EmbeddingResponse {
-    embeddings: Vec<Vec<f32>>,
-    model: String,
-    usage: Option<HashMap<String, u32>>,
-}
-
-// Ollama-specific request/response structures
-#[derive(Debug, Serialize, Deserialize)]
-struct OllamaEmbeddingRequest {
-    model: String,
-    prompt: String,
-}
-
-#[derive(Debug, Serialize, Deserialize)]
-struct OllamaEmbeddingResponse {
-    embedding: Vec<f32>,
-}
-
-pub struct LanceStore {
-    datasets: Arc<RwLock<HashMap<String, Arc<Dataset>>>>,
-    data_dir: PathBuf,
-    http_client: reqwest::Client,
-}
-
-impl LanceStore {
-    pub async fn new(data_dir: PathBuf) -> Result<Self, DBError> {
-        // Create data directory if it doesn't exist
-        std::fs::create_dir_all(&data_dir)
-            .map_err(|e| DBError(format!("Failed to create Lance data directory: {}", e)))?;
-        
-        let http_client = reqwest::Client::builder()
-            .timeout(std::time::Duration::from_secs(30))
-            .build()
-            .map_err(|e| DBError(format!("Failed to create HTTP client: {}", e)))?;
-        
-        Ok(Self {
-            datasets: Arc::new(RwLock::new(HashMap::new())),
-            data_dir,
-            http_client,
-        })
-    }
-    
-    /// Get embedding service URL from Redis config, default to local Ollama
-    async fn get_embedding_url(&self, server: &crate::server::Server) -> Result<String, DBError> {
-        // Get the embedding URL from Redis config directly from storage
-        let storage = server.current_storage()?;
-        match storage.hget("config:core:aiembed", "url")? {
-            Some(url) => Ok(url),
-            None => Ok("http://localhost:11434".to_string()), // Default to local Ollama
-        }
-    }
-    
-    /// Check if we're using Ollama (default) or custom embedding service
-    async fn is_ollama_service(&self, server: &crate::server::Server) -> Result<bool, DBError> {
-        let url = self.get_embedding_url(server).await?;
-        Ok(url.contains("localhost:11434") || url.contains("127.0.0.1:11434"))
-    }
-    
-    /// Call external embedding service (Ollama or custom)
-    async fn call_embedding_service(
-        &self,
-        server: &crate::server::Server,
-        texts: Option<Vec<String>>,
-        images: Option<Vec<String>>,
-    ) -> Result<Vec<Vec<f32>>, DBError> {
-        let base_url = self.get_embedding_url(server).await?;
-        let is_ollama = self.is_ollama_service(server).await?;
-        
-        if is_ollama {
-            // Use Ollama API format
-            if let Some(texts) = texts {
-                let mut embeddings = Vec::new();
-                for text in texts {
-                    let url = format!("{}/api/embeddings", base_url);
-                    let request = OllamaEmbeddingRequest {
-                        model: "nomic-embed-text".to_string(),
-                        prompt: text,
-                    };
-                    
-                    let response = self.http_client
-                        .post(&url)
-                        .json(&request)
-                        .send()
-                        .await
-                        .map_err(|e| DBError(format!("Failed to call Ollama embedding service: {}", e)))?;
-                    
-                    if !response.status().is_success() {
-                        let status = response.status();
-                        let error_text = response.text().await.unwrap_or_default();
-                        return Err(DBError(format!(
-                            "Ollama embedding service returned error {}: {}",
-                            status, error_text
-                        )));
-                    }
-                    
-                    let ollama_response: OllamaEmbeddingResponse = response
-                        .json()
-                        .await
-                        .map_err(|e| DBError(format!("Failed to parse Ollama embedding response: {}", e)))?;
-                    
-                    embeddings.push(ollama_response.embedding);
-                }
-                Ok(embeddings)
-            } else if let Some(_images) = images {
-                // Ollama doesn't support image embeddings with this API yet
-                Err(DBError("Image embeddings not supported with Ollama. Please configure a custom embedding service.".to_string()))
-            } else {
-                Err(DBError("No text or images provided for embedding".to_string()))
-            }
-        } else {
-            // Use custom embedding service API format
-            let request = EmbeddingRequest {
-                texts,
-                images,
-                model: None, // Let the service use its default
-            };
-            
-            let response = self.http_client
-                .post(&base_url)
-                .json(&request)
-                .send()
-                .await
-                .map_err(|e| DBError(format!("Failed to call embedding service: {}", e)))?;
-            
-            if !response.status().is_success() {
-                let status = response.status();
-                let error_text = response.text().await.unwrap_or_default();
-                return Err(DBError(format!(
-                    "Embedding service returned error {}: {}",
-                    status, error_text
-                )));
-            }
-            
-            let embedding_response: EmbeddingResponse = response
-                .json()
-                .await
-                .map_err(|e| DBError(format!("Failed to parse embedding response: {}", e)))?;
-            
-            Ok(embedding_response.embeddings)
-        }
-    }
-    
-    pub async fn embed_text(
-        &self, 
-        server: &crate::server::Server,
-        texts: Vec<String>
-    ) -> Result<Vec<Vec<f32>>, DBError> {
-        if texts.is_empty() {
-            return Ok(Vec::new());
-        }
-        
-        self.call_embedding_service(server, Some(texts), None).await
-    }
-    
-    pub async fn embed_image(
-        &self,
-        server: &crate::server::Server,
-        image_bytes: Vec<u8>
-    ) -> Result<Vec<f32>, DBError> {
-        // Convert image bytes to base64
-        let base64_image = base64::engine::general_purpose::STANDARD.encode(&image_bytes);
-        
-        let embeddings = self.call_embedding_service(
-            server, 
-            None, 
-            Some(vec![base64_image])
-        ).await?;
-        
-        embeddings.into_iter()
-            .next()
-            .ok_or_else(|| DBError("No embedding returned for image".to_string()))
-    }
-    
-    pub async fn create_dataset(
-        &self,
-        name: &str,
-        schema: Schema,
-    ) -> Result<(), DBError> {
-        let dataset_path = self.data_dir.join(format!("{}.lance", name));
-        
-        // Create empty dataset with schema
-        let write_params = WriteParams {
-            mode: WriteMode::Create,
-            ..Default::default()
-        };
-        
-        // Create an empty RecordBatch with the schema
-        let empty_batch = RecordBatch::new_empty(Arc::new(schema));
-        
-        // Use RecordBatchReader for Lance 0.33
-        let reader = VecRecordBatchReader::new(vec![empty_batch]);
-        let dataset = Dataset::write(
-            reader,
-            dataset_path.to_str().unwrap(),
-            Some(write_params)
-        ).await
-        .map_err(|e| DBError(format!("Failed to create dataset: {}", e)))?;
-        
-        let mut datasets = self.datasets.write().await;
-        datasets.insert(name.to_string(), Arc::new(dataset));
-        
-        Ok(())
-    }
-    
-    pub async fn write_vectors(
-        &self,
-        dataset_name: &str,
-        vectors: Vec<Vec<f32>>,
-        metadata: Option<HashMap<String, Vec<String>>>,
-    ) -> Result<usize, DBError> {
-        let dataset_path = self.data_dir.join(format!("{}.lance", dataset_name));
-        
-        // Open or get cached dataset
-        let _dataset = self.get_or_open_dataset(dataset_name).await?;
-        
-        // Build RecordBatch
-        let num_vectors = vectors.len();
-        if num_vectors == 0 {
-            return Ok(0);
-        }
-        
-        let dim = vectors.first()
-            .ok_or_else(|| DBError("Empty vectors".to_string()))?
-            .len();
-        
-        // Flatten vectors
-        let flat_vectors: Vec<f32> = vectors.into_iter().flatten().collect();
-        let values_array = Float32Array::from(flat_vectors);
-        let field = Arc::new(Field::new("item", DataType::Float32, true));
-        let vector_array = FixedSizeListArray::try_new(
-            field,
-            dim as i32,
-            Arc::new(values_array),
-            None
-        ).map_err(|e| DBError(format!("Failed to create vector array: {}", e)))?;
-        
-        let mut arrays: Vec<ArrayRef> = vec![Arc::new(vector_array)];
-        let mut fields = vec![Field::new(
-            "vector",
-            DataType::FixedSizeList(
-                Arc::new(Field::new("item", DataType::Float32, true)),
-                dim as i32
-            ),
-            false
-        )];
-        
-        // Add metadata columns if provided
-        if let Some(metadata) = metadata {
-            for (key, values) in metadata {
-                if values.len() != num_vectors {
-                    return Err(DBError(format!(
-                        "Metadata field '{}' has {} values but expected {}", 
-                        key, values.len(), num_vectors
-                    )));
-                }
-                let array = StringArray::from(values);
-                arrays.push(Arc::new(array));
-                fields.push(Field::new(&key, DataType::Utf8, true));
-            }
-        }
-        
-        let schema = Arc::new(Schema::new(fields));
-        let batch = RecordBatch::try_new(schema, arrays)
-            .map_err(|e| DBError(format!("Failed to create RecordBatch: {}", e)))?;
-        
-        // Append to dataset
-        let write_params = WriteParams {
-            mode: WriteMode::Append,
-            ..Default::default()
-        };
-        
-        let reader = VecRecordBatchReader::new(vec![batch]);
-        Dataset::write(
-            reader,
-            dataset_path.to_str().unwrap(),
-            Some(write_params)
-        ).await
-        .map_err(|e| DBError(format!("Failed to write to dataset: {}", e)))?;
-        
-        // Refresh cached dataset
-        let mut datasets = self.datasets.write().await;
-        datasets.remove(dataset_name);
-        
-        Ok(num_vectors)
-    }
-    
-    pub async fn search_vectors(
-        &self,
-        dataset_name: &str,
-        query_vector: Vec<f32>,
-        k: usize,
-        nprobes: Option<usize>,
-        _refine_factor: Option<usize>,
-    ) -> Result<Vec<(f32, HashMap<String, String>)>, DBError> {
-        let dataset = self.get_or_open_dataset(dataset_name).await?;
-        
-        // Build query
-        let query_array = Float32Array::from(query_vector.clone());
-        let mut query = dataset.scan();
-        query.nearest(
-            "vector",
-            &query_array,
-            k,
-        ).map_err(|e| DBError(format!("Failed to build search query: {}", e)))?;
-        
-        if let Some(nprobes) = nprobes {
-            query.nprobs(nprobes);
-        }
-        
-        // Note: refine_factor might not be available in this Lance version
-        // if let Some(refine) = refine_factor {
-        //     query.refine_factor(refine);
-        // }
-        
-        // Execute search
-        let results = query
-            .try_into_stream()
-            .await
-            .map_err(|e| DBError(format!("Failed to execute search: {}", e)))?
-            .try_collect::<Vec<_>>()
-            .await
-            .map_err(|e| DBError(format!("Failed to collect results: {}", e)))?;
-        
-        // Process results
-        let mut output = Vec::new();
-        for batch in results {
-            // Get distances
-            let distances = batch
-                .column_by_name("_distance")
-                .ok_or_else(|| DBError("No distance column".to_string()))?
-                .as_any()
-                .downcast_ref::<Float32Array>()
-                .ok_or_else(|| DBError("Invalid distance type".to_string()))?;
-            
-            // Get metadata
-            for i in 0..batch.num_rows() {
-                let distance = distances.value(i);
-                let mut metadata = HashMap::new();
-                
-                for field in batch.schema().fields() {
-                    if field.name() != "vector" && field.name() != "_distance" {
-                        if let Some(col) = batch.column_by_name(field.name()) {
-                            if let Some(str_array) = col.as_any().downcast_ref::<StringArray>() {
-                                if !str_array.is_null(i) {
-                                    metadata.insert(
-                                        field.name().to_string(),
-                                        str_array.value(i).to_string()
-                                    );
-                                }
-                            }
-                        }
-                    }
-                }
-                
-                output.push((distance, metadata));
-            }
-        }
-        
-        Ok(output)
-    }
-    
-    pub async fn store_multimodal(
-        &self,
-        server: &crate::server::Server,
-        dataset_name: &str,
-        text: Option<String>,
-        image_bytes: Option<Vec<u8>>,
-        metadata: HashMap<String, String>,
-    ) -> Result<String, DBError> {
-        // Generate ID
-        let id = uuid::Uuid::new_v4().to_string();
-        
-        // Generate embeddings using external service
-        let embedding = if let Some(text) = text.as_ref() {
-            self.embed_text(server, vec![text.clone()]).await?
-                .into_iter()
-                .next()
-                .ok_or_else(|| DBError("No embedding returned".to_string()))?
-        } else if let Some(img) = image_bytes.as_ref() {
-            self.embed_image(server, img.clone()).await?
-        } else {
-            return Err(DBError("No text or image provided".to_string()));
-        };
-        
-        // Prepare metadata
-        let mut full_metadata = metadata;
-        full_metadata.insert("id".to_string(), id.clone());
-        if let Some(text) = text {
-            full_metadata.insert("text".to_string(), text);
-        }
-        if let Some(img) = image_bytes {
-            full_metadata.insert("image_base64".to_string(), base64::engine::general_purpose::STANDARD.encode(img));
-        }
-        
-        // Convert metadata to column vectors
-        let mut metadata_cols = HashMap::new();
-        for (key, value) in full_metadata {
-            metadata_cols.insert(key, vec![value]);
-        }
-        
-        // Write to dataset
-        self.write_vectors(dataset_name, vec![embedding], Some(metadata_cols)).await?;
-        
-        Ok(id)
-    }
-    
-    pub async fn search_with_text(
-        &self,
-        server: &crate::server::Server,
-        dataset_name: &str,
-        query_text: String,
-        k: usize,
-        nprobes: Option<usize>,
-        refine_factor: Option<usize>,
-    ) -> Result<Vec<(f32, HashMap<String, String>)>, DBError> {
-        // Embed the query text using external service
-        let embeddings = self.embed_text(server, vec![query_text]).await?;
-        let query_vector = embeddings.into_iter()
-            .next()
-            .ok_or_else(|| DBError("No embedding returned for query".to_string()))?;
-        
-        // Search with the embedding
-        self.search_vectors(dataset_name, query_vector, k, nprobes, refine_factor).await
-    }
-    
-    pub async fn create_index(
-        &self,
-        dataset_name: &str,
-        index_type: &str,
-        num_partitions: Option<usize>,
-        num_sub_vectors: Option<usize>,
-    ) -> Result<(), DBError> {
-        let _dataset = self.get_or_open_dataset(dataset_name).await?;
-        
-        match index_type.to_uppercase().as_str() {
-            "IVF_PQ" => {
-                let ivf_params = IvfBuildParams {
-                    num_partitions: num_partitions.unwrap_or(256),
-                    ..Default::default()
-                };
-                let pq_params = PQBuildParams {
-                    num_sub_vectors: num_sub_vectors.unwrap_or(16),
-                    ..Default::default()
-                };
-                let params = VectorIndexParams::with_ivf_pq_params(
-                    MetricType::L2,
-                    ivf_params,
-                    pq_params,
-                );
-                
-                // Get a mutable reference to the dataset
-                let mut dataset_mut = Dataset::open(self.data_dir.join(format!("{}.lance", dataset_name)).to_str().unwrap())
-                    .await
-                    .map_err(|e| DBError(format!("Failed to open dataset for indexing: {}", e)))?;
-                
-                dataset_mut.create_index(
-                    &["vector"],
-                    lance_index::IndexType::Vector,
-                    None,
-                    &params,
-                    true
-                ).await
-                .map_err(|e| DBError(format!("Failed to create index: {}", e)))?;
-            }
-            _ => return Err(DBError(format!("Unsupported index type: {}", index_type))),
-        }
-        
-        Ok(())
-    }
-    
-    async fn get_or_open_dataset(&self, name: &str) -> Result<Arc<Dataset>, DBError> {
-        let mut datasets = self.datasets.write().await;
-        
-        if let Some(dataset) = datasets.get(name) {
-            return Ok(dataset.clone());
-        }
-        
-        let dataset_path = self.data_dir.join(format!("{}.lance", name));
-        if !dataset_path.exists() {
-            return Err(DBError(format!("Dataset '{}' does not exist", name)));
-        }
-        
-        let dataset = Dataset::open(dataset_path.to_str().unwrap())
-            .await
-            .map_err(|e| DBError(format!("Failed to open dataset: {}", e)))?;
-        
-        let dataset = Arc::new(dataset);
-        datasets.insert(name.to_string(), dataset.clone());
-        
-        Ok(dataset)
-    }
-    
-    pub async fn list_datasets(&self) -> Result<Vec<String>, DBError> {
-        let mut datasets = Vec::new();
-        
-        let entries = std::fs::read_dir(&self.data_dir)
-            .map_err(|e| DBError(format!("Failed to read data directory: {}", e)))?;
-        
-        for entry in entries {
-            let entry = entry.map_err(|e| DBError(format!("Failed to read entry: {}", e)))?;
-            let path = entry.path();
-            
-            if path.is_dir() {
-                if let Some(name) = path.file_name() {
-                    if let Some(name_str) = name.to_str() {
-                        if name_str.ends_with(".lance") {
-                            let dataset_name = name_str.trim_end_matches(".lance");
-                            datasets.push(dataset_name.to_string());
-                        }
-                    }
-                }
-            }
-        }
-        
-        Ok(datasets)
-    }
-    
-    pub async fn drop_dataset(&self, name: &str) -> Result<(), DBError> {
-        // Remove from cache
-        let mut datasets = self.datasets.write().await;
-        datasets.remove(name);
-        
-        // Delete from disk
-        let dataset_path = self.data_dir.join(format!("{}.lance", name));
-        if dataset_path.exists() {
-            std::fs::remove_dir_all(dataset_path)
-                .map_err(|e| DBError(format!("Failed to delete dataset: {}", e)))?;
-        }
-        
-        Ok(())
-    }
-    
-    pub async fn get_dataset_info(&self, name: &str) -> Result<HashMap<String, String>, DBError> {
-        let dataset = self.get_or_open_dataset(name).await?;
-        
-        let mut info = HashMap::new();
-        info.insert("name".to_string(), name.to_string());
-        info.insert("version".to_string(), dataset.version().version.to_string());
-        info.insert("num_rows".to_string(), dataset.count_rows(None).await?.to_string());
-        
-        // Get schema info
-        let schema = dataset.schema();
-        let fields: Vec<String> = schema.fields
-            .iter()
-            .map(|f| format!("{}:{}", f.name, f.data_type()))
-            .collect();
-        info.insert("schema".to_string(), fields.join(", "));
-        
-        Ok(info)
-    }
-}

src/lib.rs

@@ -1,11 +1,16 @@
-pub mod age;   // NEW
+pub mod age;
+pub mod sym;
 pub mod cmd;
 pub mod crypto;
 pub mod error;
-pub mod lance_store;    // Add Lance store module
 pub mod options;
 pub mod protocol;
+pub mod rpc;
+pub mod rpc_server;
 pub mod server;
 pub mod storage;
-pub mod storage_trait;  // Add this
-pub mod storage_sled;   // Add this
+pub mod storage_trait;
+pub mod storage_sled;
+pub mod admin_meta;
+pub mod tantivy_search;
+pub mod search_cmd;

src/main.rs

@@ -3,6 +3,7 @@
 use tokio::net::TcpListener;
 
 use herodb::server;
+use herodb::rpc_server;
 
 use clap::Parser;
 
@@ -22,18 +23,29 @@ struct Args {
     #[arg(long)]
     debug: bool,
 
-
-    /// Master encryption key for encrypted databases
+    /// Master encryption key for encrypted databases (deprecated; ignored for data DBs)
     #[arg(long)]
     encryption_key: Option<String>,
 
-    /// Encrypt the database
+    /// Encrypt the database (deprecated; ignored for data DBs)
     #[arg(long)]
     encrypt: bool,
 
+    /// Enable RPC management server
+    #[arg(long)]
+    enable_rpc: bool,
+
+    /// RPC server port (default: 8080)
+    #[arg(long, default_value = "8080")]
+    rpc_port: u16,
+
     /// Use the sled backend
     #[arg(long)]
     sled: bool,
+
+    /// Admin secret used to encrypt DB 0 and authorize admin access (required)
+    #[arg(long)]
+    admin_secret: String,
 }
 
 #[tokio::main]
@@ -48,9 +60,19 @@ async fn main() {
         .await
         .unwrap();
 
+    // deprecation warnings for legacy flags
+    if args.encrypt || args.encryption_key.is_some() {
+        eprintln!("warning: --encrypt and --encryption-key are deprecated and ignored for data DBs. Admin DB 0 is always encrypted with --admin-secret.");
+    }
+    // basic validation for admin secret
+    if args.admin_secret.trim().is_empty() {
+        eprintln!("error: --admin-secret must not be empty");
+        std::process::exit(2);
+    }
+
     // new DB option
     let option = herodb::options::DBOption {
-        dir: args.dir,
+        dir: args.dir.clone(),
         port,
         debug: args.debug,
         encryption_key: args.encryption_key,
@@ -60,14 +82,42 @@ async fn main() {
         } else {
             herodb::options::BackendType::Redb
         },
+        admin_secret: args.admin_secret.clone(),
     };
 
+    let backend = option.backend.clone();
+
+    // Bootstrap admin DB 0 before opening any server storage
+    if let Err(e) = herodb::admin_meta::ensure_bootstrap(&args.dir, backend.clone(), &args.admin_secret) {
+        eprintln!("Failed to bootstrap admin DB 0: {}", e.0);
+        std::process::exit(2);
+    }
+
     // new server
     let server = server::Server::new(option).await;
 
     // Add a small delay to ensure the port is ready
     tokio::time::sleep(std::time::Duration::from_millis(100)).await;
 
+    // Start RPC server if enabled
+    let _rpc_handle = if args.enable_rpc {
+        let rpc_addr = format!("127.0.0.1:{}", args.rpc_port).parse().unwrap();
+        let base_dir = args.dir.clone();
+
+        match rpc_server::start_rpc_server(rpc_addr, base_dir, backend, args.admin_secret.clone()).await {
+            Ok(handle) => {
+                println!("RPC management server started on port {}", args.rpc_port);
+                Some(handle)
+            }
+            Err(e) => {
+                eprintln!("Failed to start RPC server: {}", e);
+                None
+            }
+        }
+    } else {
+        None
+    };
+
     // accept new connections
     loop {
         let stream = listener.accept().await;

src/options.rs

@@ -1,7 +1,8 @@
-#[derive(Debug, Clone)]
+#[derive(Debug, Clone, PartialEq, Eq)]
 pub enum BackendType {
     Redb,
     Sled,
+    Tantivy, // Full-text search backend (no KV storage)
 }
 
 #[derive(Debug, Clone)]
@@ -9,7 +10,11 @@ pub struct DBOption {
     pub dir: String,
     pub port: u16,
     pub debug: bool,
+    // Deprecated for data DBs; retained for backward-compat on CLI parsing
     pub encrypt: bool,
+    // Deprecated for data DBs; retained for backward-compat on CLI parsing
     pub encryption_key: Option<String>,
     pub backend: BackendType,
+    // New: required admin secret, used to encrypt DB 0 and authorize admin operations
+    pub admin_secret: String,
 }

src/rpc.rs

@@ -0,0 +1,646 @@
+use std::collections::HashMap;
+use std::sync::Arc;
+use tokio::sync::RwLock;
+use jsonrpsee::{core::RpcResult, proc_macros::rpc};
+use serde::{Deserialize, Serialize};
+use sha2::{Digest, Sha256};
+
+use crate::server::Server;
+use crate::options::DBOption;
+use crate::admin_meta;
+
+/// Database backend types
+#[derive(Debug, Clone, Serialize, Deserialize)]
+pub enum BackendType {
+    Redb,
+    Sled,
+    Tantivy, // Full-text search backend (no KV storage)
+    // Future: InMemory, Custom(String)
+}
+
+/// Database configuration
+#[derive(Debug, Clone, Serialize, Deserialize)]
+pub struct DatabaseConfig {
+    pub name: Option<String>,
+    pub storage_path: Option<String>,
+    pub max_size: Option<u64>,
+    pub redis_version: Option<String>,
+}
+
+/// Database information returned by metadata queries
+#[derive(Debug, Clone, Serialize, Deserialize)]
+pub struct DatabaseInfo {
+    pub id: u64,
+    pub name: Option<String>,
+    pub backend: BackendType,
+    pub encrypted: bool,
+    pub redis_version: Option<String>,
+    pub storage_path: Option<String>,
+    pub size_on_disk: Option<u64>,
+    pub key_count: Option<u64>,
+    pub created_at: u64,
+    pub last_access: Option<u64>,
+}
+
+/// Access permissions for database keys
+#[derive(Debug, Clone, Serialize, Deserialize, PartialEq)]
+pub enum Permissions {
+    Read,
+    ReadWrite,
+}
+
+
+
+/// Access key information returned by RPC
+#[derive(Debug, Clone, Serialize, Deserialize)]
+pub struct AccessKeyInfo {
+    pub hash: String,
+    pub permissions: Permissions,
+    pub created_at: u64,
+}
+
+/// Hash a plaintext key using SHA-256
+pub fn hash_key(key: &str) -> String {
+    let mut hasher = Sha256::new();
+    hasher.update(key.as_bytes());
+    format!("{:x}", hasher.finalize())
+}
+
+/// RPC trait for HeroDB management
+#[rpc(server, client, namespace = "herodb")]
+pub trait Rpc {
+    /// Create a new database with specified configuration
+    #[method(name = "createDatabase")]
+    async fn create_database(
+        &self,
+        backend: BackendType,
+        config: DatabaseConfig,
+        encryption_key: Option<String>,
+    ) -> RpcResult<u64>;
+
+    /// Set encryption for an existing database (write-only key)
+    #[method(name = "setEncryption")]
+    async fn set_encryption(&self, db_id: u64, encryption_key: String) -> RpcResult<bool>;
+
+    /// List all managed databases
+    #[method(name = "listDatabases")]
+    async fn list_databases(&self) -> RpcResult<Vec<DatabaseInfo>>;
+
+    /// Get detailed information about a specific database
+    #[method(name = "getDatabaseInfo")]
+    async fn get_database_info(&self, db_id: u64) -> RpcResult<DatabaseInfo>;
+
+    /// Delete a database
+    #[method(name = "deleteDatabase")]
+    async fn delete_database(&self, db_id: u64) -> RpcResult<bool>;
+
+    /// Get server statistics
+    #[method(name = "getServerStats")]
+    async fn get_server_stats(&self) -> RpcResult<HashMap<String, serde_json::Value>>;
+
+    /// Add an access key to a database
+    #[method(name = "addAccessKey")]
+    async fn add_access_key(&self, db_id: u64, key: String, permissions: String) -> RpcResult<bool>;
+
+    /// Delete an access key from a database
+    #[method(name = "deleteAccessKey")]
+    async fn delete_access_key(&self, db_id: u64, key_hash: String) -> RpcResult<bool>;
+
+    /// List all access keys for a database
+    #[method(name = "listAccessKeys")]
+    async fn list_access_keys(&self, db_id: u64) -> RpcResult<Vec<AccessKeyInfo>>;
+
+    /// Set database public/private status
+    #[method(name = "setDatabasePublic")]
+    async fn set_database_public(&self, db_id: u64, public: bool) -> RpcResult<bool>;
+
+    // ----- Full-text (Tantivy) minimal RPC endpoints -----
+
+    /// Create a new FT index in a Tantivy-backed DB
+    #[method(name = "ftCreate")]
+    async fn ft_create(
+        &self,
+        db_id: u64,
+        index_name: String,
+        schema: Vec<(String, String, Vec<String>)>,
+    ) -> RpcResult<bool>;
+
+    /// Add or replace a document in an FT index
+    #[method(name = "ftAdd")]
+    async fn ft_add(
+        &self,
+        db_id: u64,
+        index_name: String,
+        doc_id: String,
+        score: f64,
+        fields: HashMap<String, String>,
+    ) -> RpcResult<bool>;
+
+    /// Search an FT index
+    #[method(name = "ftSearch")]
+    async fn ft_search(
+        &self,
+        db_id: u64,
+        index_name: String,
+        query: String,
+        filters: Option<Vec<(String, String)>>,
+        limit: Option<usize>,
+        offset: Option<usize>,
+        return_fields: Option<Vec<String>>,
+    ) -> RpcResult<serde_json::Value>;
+
+    /// Delete a document by id from an FT index
+    #[method(name = "ftDel")]
+    async fn ft_del(&self, db_id: u64, index_name: String, doc_id: String) -> RpcResult<bool>;
+
+    /// Get FT index info
+    #[method(name = "ftInfo")]
+    async fn ft_info(&self, db_id: u64, index_name: String) -> RpcResult<serde_json::Value>;
+
+    /// Drop an FT index
+    #[method(name = "ftDrop")]
+    async fn ft_drop(&self, db_id: u64, index_name: String) -> RpcResult<bool>;
+}
+
+/// RPC Server implementation
+pub struct RpcServerImpl {
+    /// Base directory for database files
+    base_dir: String,
+    /// Managed database servers
+    servers: Arc<RwLock<HashMap<u64, Arc<Server>>>>,
+    /// Default backend type
+    backend: crate::options::BackendType,
+    /// Admin secret used to encrypt DB 0 and authorize admin access
+    admin_secret: String,
+}
+
+impl RpcServerImpl {
+    /// Create a new RPC server instance
+    pub fn new(base_dir: String, backend: crate::options::BackendType, admin_secret: String) -> Self {
+        Self {
+            base_dir,
+            servers: Arc::new(RwLock::new(HashMap::new())),
+            backend,
+            admin_secret,
+        }
+    }
+
+    /// Get or create a server instance for the given database ID
+    async fn get_or_create_server(&self, db_id: u64) -> Result<Arc<Server>, jsonrpsee::types::ErrorObjectOwned> {
+        // Check if server already exists
+        {
+            let servers = self.servers.read().await;
+            if let Some(server) = servers.get(&db_id) {
+                return Ok(server.clone());
+            }
+        }
+
+        // Validate existence via admin DB 0 (metadata), not filesystem presence
+        let exists = admin_meta::db_exists(&self.base_dir, self.backend.clone(), &self.admin_secret, db_id)
+            .map_err(|e| jsonrpsee::types::ErrorObjectOwned::owned(-32000, e.0, None::<()>))?;
+        if !exists {
+            return Err(jsonrpsee::types::ErrorObjectOwned::owned(
+                -32000,
+                format!("Database {} not found", db_id),
+                None::<()>
+            ));
+        }
+
+        // Resolve effective backend for this db from admin meta or filesystem; fallback to default
+        let meta_backend = admin_meta::get_database_backend(&self.base_dir, self.backend.clone(), &self.admin_secret, db_id)
+            .ok()
+            .flatten();
+        let db_path = std::path::PathBuf::from(&self.base_dir).join(format!("{}.db", db_id));
+        let sniffed_backend = if db_path.exists() {
+            if db_path.is_file() {
+                Some(crate::options::BackendType::Redb)
+            } else if db_path.is_dir() {
+                Some(crate::options::BackendType::Sled)
+            } else {
+                None
+            }
+        } else {
+            None
+        };
+        let effective_backend = meta_backend.clone().or(sniffed_backend).unwrap_or(self.backend.clone());
+        if effective_backend != self.backend {
+            eprintln!(
+                "notice: get_or_create_server: db {} backend resolved to {:?} (server default {:?})",
+                db_id, effective_backend, self.backend
+            );
+        }
+        // If we had to sniff (no meta), persist the resolved backend
+        if meta_backend.is_none() {
+            let _ = admin_meta::set_database_backend(&self.base_dir, self.backend.clone(), &self.admin_secret, db_id, effective_backend.clone());
+        }
+
+        // Create server instance with resolved backend
+        let is_tantivy = matches!(effective_backend, crate::options::BackendType::Tantivy);
+        let db_option = DBOption {
+            dir: self.base_dir.clone(),
+            port: 0, // Not used for RPC-managed databases
+            debug: false,
+            encryption_key: None,
+            encrypt: false,
+            backend: effective_backend.clone(),
+            admin_secret: self.admin_secret.clone(),
+        };
+
+        let mut server = Server::new(db_option).await;
+
+        // Set the selected database to the db_id
+        server.selected_db = db_id;
+
+        // Lazily open/create physical storage according to admin meta (per-db encryption)
+        // Skip for Tantivy backend (no KV storage to open)
+        if !is_tantivy {
+            let _ = server.current_storage();
+        }
+
+        // Store the server
+        let mut servers = self.servers.write().await;
+        servers.insert(db_id, Arc::new(server.clone()));
+
+        Ok(Arc::new(server))
+    }
+
+    /// Discover existing database IDs from admin DB 0
+    async fn discover_databases(&self) -> Vec<u64> {
+        admin_meta::list_dbs(&self.base_dir, self.backend.clone(), &self.admin_secret)
+            .unwrap_or_default()
+    }
+
+
+
+
+
+
+    /// Build database file path for given server/db_id
+    fn db_file_path(&self, server: &Server, db_id: u64) -> std::path::PathBuf {
+        std::path::PathBuf::from(&server.option.dir).join(format!("{}.db", db_id))
+    }
+
+    /// Recursively compute size on disk for the database path
+    fn compute_size_on_disk(&self, path: &std::path::Path) -> Option<u64> {
+        fn dir_size(p: &std::path::Path) -> u64 {
+            if p.is_file() {
+                std::fs::metadata(p).map(|m| m.len()).unwrap_or(0)
+            } else if p.is_dir() {
+                let mut total = 0u64;
+                if let Ok(read) = std::fs::read_dir(p) {
+                    for entry in read.flatten() {
+                        total += dir_size(&entry.path());
+                    }
+                }
+                total
+            } else {
+                0
+            }
+        }
+        Some(dir_size(path))
+    }
+
+    /// Extract created and last access times (secs) from a path, with fallbacks
+    fn get_file_times_secs(path: &std::path::Path) -> (u64, Option<u64>) {
+        let now = std::time::SystemTime::now();
+        let created = std::fs::metadata(path)
+            .and_then(|m| m.created().or_else(|_| m.modified()))
+            .unwrap_or(now)
+            .duration_since(std::time::UNIX_EPOCH)
+            .unwrap_or_default()
+            .as_secs();
+
+        let last_access = std::fs::metadata(path)
+            .and_then(|m| m.accessed())
+            .ok()
+            .and_then(|t| t.duration_since(std::time::UNIX_EPOCH).ok().map(|d| d.as_secs()));
+
+        (created, last_access)
+    }
+
+    /// Compose a DatabaseInfo by probing storage and filesystem, with admin meta for access key count
+    async fn build_database_info(&self, db_id: u64, server: &Server) -> DatabaseInfo {
+        // Probe storage to determine encryption state
+        let storage = server.current_storage().ok();
+        let encrypted = storage.as_ref().map(|s| s.is_encrypted()).unwrap_or(server.option.encrypt);
+
+        // Get actual key count from storage
+        let key_count = storage.as_ref()
+            .and_then(|s| s.dbsize().ok())
+            .map(|count| count as u64);
+
+        // Get database name from admin meta
+        let name = admin_meta::get_database_name(&self.base_dir, self.backend.clone(), &self.admin_secret, db_id)
+            .ok()
+            .flatten();
+
+        // Compute size on disk and timestamps from the DB file path
+        let db_path = self.db_file_path(server, db_id);
+        let size_on_disk = self.compute_size_on_disk(&db_path);
+        let (created_at, last_access) = Self::get_file_times_secs(&db_path);
+
+        let backend = match server.option.backend {
+            crate::options::BackendType::Redb => BackendType::Redb,
+            crate::options::BackendType::Sled => BackendType::Sled,
+            crate::options::BackendType::Tantivy => BackendType::Tantivy,
+        };
+
+        DatabaseInfo {
+            id: db_id,
+            name,
+            backend,
+            encrypted,
+            redis_version: Some("7.0".to_string()),
+            storage_path: Some(server.option.dir.clone()),
+            size_on_disk,
+            key_count,
+            created_at,
+            last_access,
+        }
+    }
+}
+
+#[jsonrpsee::core::async_trait]
+impl RpcServer for RpcServerImpl {
+    async fn create_database(
+        &self,
+        backend: BackendType,
+        config: DatabaseConfig,
+        encryption_key: Option<String>,
+    ) -> RpcResult<u64> {
+        // Allocate new ID via admin DB 0
+        let db_id = admin_meta::allocate_next_id(&self.base_dir, self.backend.clone(), &self.admin_secret)
+            .map_err(|e| jsonrpsee::types::ErrorObjectOwned::owned(-32000, e.0, None::<()>))?;
+
+        // Persist per-db encryption key in admin DB 0 if provided
+        if let Some(ref key) = encryption_key {
+            admin_meta::set_enc_key(&self.base_dir, self.backend.clone(), &self.admin_secret, db_id, key)
+                .map_err(|e| jsonrpsee::types::ErrorObjectOwned::owned(-32000, e.0, None::<()>))?;
+        }
+
+        // Persist database name if provided
+        if let Some(ref name) = config.name {
+            admin_meta::set_database_name(&self.base_dir, self.backend.clone(), &self.admin_secret, db_id, name)
+                .map_err(|e| jsonrpsee::types::ErrorObjectOwned::owned(-32000, e.0, None::<()>))?;
+        }
+
+        // Ensure base dir exists
+        if let Err(e) = std::fs::create_dir_all(&self.base_dir) {
+            return Err(jsonrpsee::types::ErrorObjectOwned::owned(-32000, format!("Failed to ensure base dir: {}", e), None::<()>));
+        }
+
+        // Map RPC backend to options backend and persist it in admin meta for this db id
+        let opt_backend = match backend {
+            BackendType::Redb => crate::options::BackendType::Redb,
+            BackendType::Sled => crate::options::BackendType::Sled,
+            BackendType::Tantivy => crate::options::BackendType::Tantivy,
+        };
+        admin_meta::set_database_backend(&self.base_dir, self.backend.clone(), &self.admin_secret, db_id, opt_backend.clone())
+            .map_err(|e| jsonrpsee::types::ErrorObjectOwned::owned(-32000, e.0, None::<()>))?;
+
+        // Create server instance using base_dir, chosen backend and admin secret
+        let is_tantivy_new = matches!(opt_backend, crate::options::BackendType::Tantivy);
+        let option = DBOption {
+            dir: self.base_dir.clone(),
+            port: 0, // Not used for RPC-managed databases
+            debug: false,
+            encryption_key: None,  // per-db key is stored in admin DB 0
+            encrypt: false,        // encryption decided per-db at open time
+            backend: opt_backend.clone(),
+            admin_secret: self.admin_secret.clone(),
+        };
+
+        let mut server = Server::new(option).await;
+        server.selected_db = db_id;
+
+        // Initialize storage to create physical <id>.db with proper encryption from admin meta
+        // Skip for Tantivy backend (no KV storage to initialize)
+        if !is_tantivy_new {
+            let _ = server.current_storage();
+        }
+
+        // Store the server in cache
+        let mut servers = self.servers.write().await;
+        servers.insert(db_id, Arc::new(server));
+
+        Ok(db_id)
+    }
+
+    async fn set_encryption(&self, _db_id: u64, _encryption_key: String) -> RpcResult<bool> {
+        // For now, return false as encryption can only be set during creation
+        let _servers = self.servers.read().await;
+        // TODO: Implement encryption setting for existing databases
+        Ok(false)
+    }
+
+    async fn list_databases(&self) -> RpcResult<Vec<DatabaseInfo>> {
+        let db_ids = self.discover_databases().await;
+        let mut result = Vec::new();
+
+        for db_id in db_ids {
+            if let Ok(server) = self.get_or_create_server(db_id).await {
+                // Build accurate info from storage/meta/fs
+                let info = self.build_database_info(db_id, &server).await;
+                result.push(info);
+            }
+        }
+
+        Ok(result)
+    }
+
+    async fn get_database_info(&self, db_id: u64) -> RpcResult<DatabaseInfo> {
+        let server = self.get_or_create_server(db_id).await?;
+        // Build accurate info from storage/meta/fs
+        let info = self.build_database_info(db_id, &server).await;
+        Ok(info)
+    }
+
+    async fn delete_database(&self, db_id: u64) -> RpcResult<bool> {
+        let mut servers = self.servers.write().await;
+
+        if let Some(_server) = servers.remove(&db_id) {
+            // Clean up database files
+            let db_path = std::path::PathBuf::from(&self.base_dir).join(format!("{}.db", db_id));
+            if db_path.exists() {
+                if db_path.is_dir() {
+                    std::fs::remove_dir_all(&db_path).ok();
+                } else {
+                    std::fs::remove_file(&db_path).ok();
+                }
+            }
+            Ok(true)
+        } else {
+            Ok(false)
+        }
+    }
+
+    async fn get_server_stats(&self) -> RpcResult<HashMap<String, serde_json::Value>> {
+        let db_ids = self.discover_databases().await;
+        let mut stats = HashMap::new();
+
+        
+        stats.insert("total_databases".to_string(), serde_json::json!(db_ids.len()));
+        stats.insert("uptime".to_string(), serde_json::json!(
+            std::time::SystemTime::now()
+                .duration_since(std::time::UNIX_EPOCH)
+                .unwrap()
+                .as_secs()
+        ));
+
+        Ok(stats)
+    }
+
+    // ----- Full-text (Tantivy) minimal RPC endpoints -----
+
+    async fn ft_create(
+        &self,
+        db_id: u64,
+        index_name: String,
+        schema: Vec<(String, String, Vec<String>)>,
+    ) -> RpcResult<bool> {
+        let server = self.get_or_create_server(db_id).await?;
+        if db_id == 0 {
+            return Err(jsonrpsee::types::ErrorObjectOwned::owned(-32000, "FT not allowed on DB 0", None::<()>));
+        }
+        if !matches!(server.option.backend, crate::options::BackendType::Tantivy) {
+            return Err(jsonrpsee::types::ErrorObjectOwned::owned(-32000, "DB backend is not Tantivy", None::<()>));
+        }
+        crate::search_cmd::ft_create_cmd(&*server, index_name, schema)
+            .await
+            .map_err(|e| jsonrpsee::types::ErrorObjectOwned::owned(-32000, e.0, None::<()>))?;
+        Ok(true)
+    }
+
+    async fn ft_add(
+        &self,
+        db_id: u64,
+        index_name: String,
+        doc_id: String,
+        score: f64,
+        fields: HashMap<String, String>,
+    ) -> RpcResult<bool> {
+        let server = self.get_or_create_server(db_id).await?;
+        if db_id == 0 {
+            return Err(jsonrpsee::types::ErrorObjectOwned::owned(-32000, "FT not allowed on DB 0", None::<()>));
+        }
+        if !matches!(server.option.backend, crate::options::BackendType::Tantivy) {
+            return Err(jsonrpsee::types::ErrorObjectOwned::owned(-32000, "DB backend is not Tantivy", None::<()>));
+        }
+        crate::search_cmd::ft_add_cmd(&*server, index_name, doc_id, score, fields)
+            .await
+            .map_err(|e| jsonrpsee::types::ErrorObjectOwned::owned(-32000, e.0, None::<()>))?;
+        Ok(true)
+    }
+
+    async fn ft_search(
+        &self,
+        db_id: u64,
+        index_name: String,
+        query: String,
+        filters: Option<Vec<(String, String)>>,
+        limit: Option<usize>,
+        offset: Option<usize>,
+        return_fields: Option<Vec<String>>,
+    ) -> RpcResult<serde_json::Value> {
+        let server = self.get_or_create_server(db_id).await?;
+        if db_id == 0 {
+            return Err(jsonrpsee::types::ErrorObjectOwned::owned(-32000, "FT not allowed on DB 0", None::<()>));
+        }
+        if !matches!(server.option.backend, crate::options::BackendType::Tantivy) {
+            return Err(jsonrpsee::types::ErrorObjectOwned::owned(-32000, "DB backend is not Tantivy", None::<()>));
+        }
+        let proto = crate::search_cmd::ft_search_cmd(
+            &*server,
+            index_name,
+            query,
+            filters.unwrap_or_default(),
+            limit,
+            offset,
+            return_fields,
+        )
+        .await
+        .map_err(|e| jsonrpsee::types::ErrorObjectOwned::owned(-32000, e.0, None::<()>))?;
+        Ok(serde_json::json!({ "resp": proto.encode() }))
+    }
+
+    async fn ft_del(&self, db_id: u64, index_name: String, doc_id: String) -> RpcResult<bool> {
+        let server = self.get_or_create_server(db_id).await?;
+        if db_id == 0 {
+            return Err(jsonrpsee::types::ErrorObjectOwned::owned(-32000, "FT not allowed on DB 0", None::<()>));
+        }
+        if !matches!(server.option.backend, crate::options::BackendType::Tantivy) {
+            return Err(jsonrpsee::types::ErrorObjectOwned::owned(-32000, "DB backend is not Tantivy", None::<()>));
+        }
+        crate::search_cmd::ft_del_cmd(&*server, index_name, doc_id)
+            .await
+            .map_err(|e| jsonrpsee::types::ErrorObjectOwned::owned(-32000, e.0, None::<()>))?;
+        Ok(true)
+    }
+
+    async fn ft_info(&self, db_id: u64, index_name: String) -> RpcResult<serde_json::Value> {
+        let server = self.get_or_create_server(db_id).await?;
+        if db_id == 0 {
+            return Err(jsonrpsee::types::ErrorObjectOwned::owned(-32000, "FT not allowed on DB 0", None::<()>));
+        }
+        if !matches!(server.option.backend, crate::options::BackendType::Tantivy) {
+            return Err(jsonrpsee::types::ErrorObjectOwned::owned(-32000, "DB backend is not Tantivy", None::<()>));
+        }
+        let proto = crate::search_cmd::ft_info_cmd(&*server, index_name)
+            .await
+            .map_err(|e| jsonrpsee::types::ErrorObjectOwned::owned(-32000, e.0, None::<()>))?;
+        Ok(serde_json::json!({ "resp": proto.encode() }))
+    }
+
+    async fn ft_drop(&self, db_id: u64, index_name: String) -> RpcResult<bool> {
+        let server = self.get_or_create_server(db_id).await?;
+        if db_id == 0 {
+            return Err(jsonrpsee::types::ErrorObjectOwned::owned(-32000, "FT not allowed on DB 0", None::<()>));
+        }
+        if !matches!(server.option.backend, crate::options::BackendType::Tantivy) {
+            return Err(jsonrpsee::types::ErrorObjectOwned::owned(-32000, "DB backend is not Tantivy", None::<()>));
+        }
+        crate::search_cmd::ft_drop_cmd(&*server, index_name)
+            .await
+            .map_err(|e| jsonrpsee::types::ErrorObjectOwned::owned(-32000, e.0, None::<()>))?;
+        Ok(true)
+    }
+
+    async fn add_access_key(&self, db_id: u64, key: String, permissions: String) -> RpcResult<bool> {
+        let perms = match permissions.to_lowercase().as_str() {
+            "read" => Permissions::Read,
+            "readwrite" => Permissions::ReadWrite,
+            _ => return Err(jsonrpsee::types::ErrorObjectOwned::owned(
+                -32000,
+                "Invalid permissions: use 'read' or 'readwrite'",
+                None::<()>
+            )),
+        };
+
+        admin_meta::add_access_key(&self.base_dir, self.backend.clone(), &self.admin_secret, db_id, &key, perms)
+            .map_err(|e| jsonrpsee::types::ErrorObjectOwned::owned(-32000, e.0, None::<()>))?;
+        Ok(true)
+    }
+
+    async fn delete_access_key(&self, db_id: u64, key_hash: String) -> RpcResult<bool> {
+        let ok = admin_meta::delete_access_key(&self.base_dir, self.backend.clone(), &self.admin_secret, db_id, &key_hash)
+            .map_err(|e| jsonrpsee::types::ErrorObjectOwned::owned(-32000, e.0, None::<()>))?;
+        Ok(ok)
+    }
+
+    async fn list_access_keys(&self, db_id: u64) -> RpcResult<Vec<AccessKeyInfo>> {
+        let pairs = admin_meta::list_access_keys(&self.base_dir, self.backend.clone(), &self.admin_secret, db_id)
+            .map_err(|e| jsonrpsee::types::ErrorObjectOwned::owned(-32000, e.0, None::<()>))?;
+        let keys: Vec<AccessKeyInfo> = pairs.into_iter().map(|(hash, perm, ts)| AccessKeyInfo {
+            hash,
+            permissions: perm,
+            created_at: ts,
+        }).collect();
+        Ok(keys)
+    }
+
+    async fn set_database_public(&self, db_id: u64, public: bool) -> RpcResult<bool> {
+        admin_meta::set_database_public(&self.base_dir, self.backend.clone(), &self.admin_secret, db_id, public)
+            .map_err(|e| jsonrpsee::types::ErrorObjectOwned::owned(-32000, e.0, None::<()>))?;
+        Ok(true)
+    }
+}

src/rpc_server.rs

@@ -0,0 +1,49 @@
+use std::net::SocketAddr;
+use jsonrpsee::server::{ServerBuilder, ServerHandle};
+use jsonrpsee::RpcModule;
+
+use crate::rpc::{RpcServer, RpcServerImpl};
+
+/// Start the RPC server on the specified address
+pub async fn start_rpc_server(addr: SocketAddr, base_dir: String, backend: crate::options::BackendType, admin_secret: String) -> Result<ServerHandle, Box<dyn std::error::Error + Send + Sync>> {
+    // Create the RPC server implementation
+    let rpc_impl = RpcServerImpl::new(base_dir, backend, admin_secret);
+
+    // Create the RPC module
+    let mut module = RpcModule::new(());
+    module.merge(RpcServer::into_rpc(rpc_impl))?;
+
+    // Build the server with both HTTP and WebSocket support
+    let server = ServerBuilder::default()
+        .build(addr)
+        .await?;
+
+    // Start the server
+    let handle = server.start(module);
+
+    println!("RPC server started on {}", addr);
+
+    Ok(handle)
+}
+
+#[cfg(test)]
+mod tests {
+    use super::*;
+    use std::time::Duration;
+
+    #[tokio::test]
+    async fn test_rpc_server_startup() {
+        let addr = "127.0.0.1:0".parse().unwrap(); // Use port 0 for auto-assignment
+        let base_dir = "/tmp/test_rpc".to_string();
+        let backend = crate::options::BackendType::Redb; // Default for test
+
+        let handle = start_rpc_server(addr, base_dir, backend, "test-admin".to_string()).await.unwrap();
+
+        // Give the server a moment to start
+        tokio::time::sleep(Duration::from_millis(100)).await;
+
+        // Stop the server
+        handle.stop().unwrap();
+        handle.stopped().await;
+    }
+}

src/search_cmd.rs

@@ -0,0 +1,352 @@
+use crate::{
+    error::DBError,
+    protocol::Protocol,
+    server::Server,
+    tantivy_search::{
+        FieldDef, Filter, FilterType, IndexConfig, NumericType, SearchOptions, TantivySearch,
+    },
+};
+use std::collections::HashMap;
+use std::sync::Arc;
+
+pub async fn ft_create_cmd(
+    server: &Server,
+    index_name: String,
+    schema: Vec<(String, String, Vec<String>)>,
+) -> Result<Protocol, DBError> {
+    if server.selected_db == 0 {
+        return Ok(Protocol::err("FT commands are not allowed on DB 0"));
+    }
+    // Enforce Tantivy backend for selected DB
+    let is_tantivy = crate::admin_meta::get_database_backend(
+        &server.option.dir,
+        server.option.backend.clone(),
+        &server.option.admin_secret,
+        server.selected_db,
+    )
+    .ok()
+    .flatten()
+    .map(|b| matches!(b, crate::options::BackendType::Tantivy))
+    .unwrap_or(false);
+    if !is_tantivy {
+        return Ok(Protocol::err("ERR DB backend is not Tantivy; FT.* commands are not allowed"));
+    }
+
+    // Parse schema into field definitions
+    let mut field_definitions = Vec::new();
+    for (field_name, field_type, options) in schema {
+        let field_def = match field_type.to_uppercase().as_str() {
+            "TEXT" => {
+                let mut sortable = false;
+                let mut no_index = false;
+                // Weight is not used in current implementation
+                let mut _weight = 1.0f32;
+                let mut i = 0;
+                while i < options.len() {
+                    match options[i].to_uppercase().as_str() {
+                        "WEIGHT" => {
+                            if i + 1 < options.len() {
+                                _weight = options[i + 1].parse::<f32>().unwrap_or(1.0);
+                                i += 2;
+                                continue;
+                            }
+                        }
+                        "SORTABLE" => {
+                            sortable = true;
+                        }
+                        "NOINDEX" => {
+                            no_index = true;
+                        }
+                        _ => {}
+                    }
+                    i += 1;
+                }
+                FieldDef::Text {
+                    stored: true,
+                    indexed: !no_index,
+                    tokenized: true,
+                    fast: sortable,
+                }
+            }
+            "NUMERIC" => {
+                // default to F64
+                let mut sortable = false;
+                for opt in &options {
+                    if opt.to_uppercase() == "SORTABLE" {
+                        sortable = true;
+                    }
+                }
+                FieldDef::Numeric {
+                    stored: true,
+                    indexed: true,
+                    fast: sortable,
+                    precision: NumericType::F64,
+                }
+            }
+            "TAG" => {
+                let mut separator = ",".to_string();
+                let mut case_sensitive = false;
+                let mut i = 0;
+                while i < options.len() {
+                    match options[i].to_uppercase().as_str() {
+                        "SEPARATOR" => {
+                            if i + 1 < options.len() {
+                                separator = options[i + 1].clone();
+                                i += 2;
+                                continue;
+                            }
+                        }
+                        "CASESENSITIVE" => {
+                            case_sensitive = true;
+                        }
+                        _ => {}
+                    }
+                    i += 1;
+                }
+                FieldDef::Tag {
+                    stored: true,
+                    separator,
+                    case_sensitive,
+                }
+            }
+            "GEO" => FieldDef::Geo { stored: true },
+            _ => {
+                return Err(DBError(format!("Unknown field type: {}", field_type)));
+            }
+        };
+        field_definitions.push((field_name, field_def));
+    }
+
+    // Create the search index
+    let search_path = server.search_index_path();
+    let config = IndexConfig::default();
+    let search_index = TantivySearch::new_with_schema(
+        search_path,
+        index_name.clone(),
+        field_definitions,
+        Some(config),
+    )?;
+
+    // Store in registry
+    let mut indexes = server.search_indexes.write().unwrap();
+    indexes.insert(index_name, Arc::new(search_index));
+
+    Ok(Protocol::SimpleString("OK".to_string()))
+}
+
+pub async fn ft_add_cmd(
+    server: &Server,
+    index_name: String,
+    doc_id: String,
+    _score: f64,
+    fields: HashMap<String, String>,
+) -> Result<Protocol, DBError> {
+    if server.selected_db == 0 {
+        return Ok(Protocol::err("FT commands are not allowed on DB 0"));
+    }
+    // Enforce Tantivy backend for selected DB
+    let is_tantivy = crate::admin_meta::get_database_backend(
+        &server.option.dir,
+        server.option.backend.clone(),
+        &server.option.admin_secret,
+        server.selected_db,
+    )
+    .ok()
+    .flatten()
+    .map(|b| matches!(b, crate::options::BackendType::Tantivy))
+    .unwrap_or(false);
+    if !is_tantivy {
+        return Ok(Protocol::err("ERR DB backend is not Tantivy; FT.* commands are not allowed"));
+    }
+    let indexes = server.search_indexes.read().unwrap();
+    let search_index = indexes
+        .get(&index_name)
+        .ok_or_else(|| DBError(format!("Index '{}' not found", index_name)))?;
+    search_index.add_document_with_fields(&doc_id, fields)?;
+    Ok(Protocol::SimpleString("OK".to_string()))
+}
+
+pub async fn ft_search_cmd(
+    server: &Server,
+    index_name: String,
+    query: String,
+    filters: Vec<(String, String)>,
+    limit: Option<usize>,
+    offset: Option<usize>,
+    return_fields: Option<Vec<String>>,
+) -> Result<Protocol, DBError> {
+    if server.selected_db == 0 {
+        return Ok(Protocol::err("FT commands are not allowed on DB 0"));
+    }
+    // Enforce Tantivy backend for selected DB
+    let is_tantivy = crate::admin_meta::get_database_backend(
+        &server.option.dir,
+        server.option.backend.clone(),
+        &server.option.admin_secret,
+        server.selected_db,
+    )
+    .ok()
+    .flatten()
+    .map(|b| matches!(b, crate::options::BackendType::Tantivy))
+    .unwrap_or(false);
+    if !is_tantivy {
+        return Ok(Protocol::err("ERR DB backend is not Tantivy; FT.* commands are not allowed"));
+    }
+    let indexes = server.search_indexes.read().unwrap();
+    let search_index = indexes
+        .get(&index_name)
+        .ok_or_else(|| DBError(format!("Index '{}' not found", index_name)))?;
+
+    let search_filters = filters
+        .into_iter()
+        .map(|(field, value)| Filter {
+            field,
+            filter_type: FilterType::Equals(value),
+        })
+        .collect();
+
+    let options = SearchOptions {
+        limit: limit.unwrap_or(10),
+        offset: offset.unwrap_or(0),
+        filters: search_filters,
+        sort_by: None,
+        return_fields,
+        highlight: false,
+    };
+
+    let results = search_index.search_with_options(&query, options)?;
+
+    // Format results as Redis protocol
+    let mut response = Vec::new();
+    // First element is the total count
+    response.push(Protocol::SimpleString(results.total.to_string()));
+    // Then each document
+    for doc in results.documents {
+        let mut doc_array = Vec::new();
+        // Add document ID if it exists
+        if let Some(id) = doc.fields.get("_id") {
+            doc_array.push(Protocol::BulkString(id.clone()));
+        }
+        // Add score
+        doc_array.push(Protocol::BulkString(doc.score.to_string()));
+        // Add fields as key-value pairs
+        for (field_name, field_value) in doc.fields {
+            if field_name != "_id" {
+                doc_array.push(Protocol::BulkString(field_name));
+                doc_array.push(Protocol::BulkString(field_value));
+            }
+        }
+        response.push(Protocol::Array(doc_array));
+    }
+
+    Ok(Protocol::Array(response))
+}
+
+pub async fn ft_del_cmd(
+    server: &Server,
+    index_name: String,
+    doc_id: String,
+) -> Result<Protocol, DBError> {
+    if server.selected_db == 0 {
+        return Ok(Protocol::err("FT commands are not allowed on DB 0"));
+    }
+    // Enforce Tantivy backend for selected DB
+    let is_tantivy = crate::admin_meta::get_database_backend(
+        &server.option.dir,
+        server.option.backend.clone(),
+        &server.option.admin_secret,
+        server.selected_db,
+    )
+    .ok()
+    .flatten()
+    .map(|b| matches!(b, crate::options::BackendType::Tantivy))
+    .unwrap_or(false);
+    if !is_tantivy {
+        return Ok(Protocol::err("ERR DB backend is not Tantivy; FT.* commands are not allowed"));
+    }
+    let indexes = server.search_indexes.read().unwrap();
+    let _search_index = indexes
+        .get(&index_name)
+        .ok_or_else(|| DBError(format!("Index '{}' not found", index_name)))?;
+    // Not fully implemented yet: Tantivy delete by term would require a writer session and commit coordination.
+    println!("Deleting document '{}' from index '{}'", doc_id, index_name);
+    Ok(Protocol::SimpleString("1".to_string()))
+}
+
+pub async fn ft_info_cmd(server: &Server, index_name: String) -> Result<Protocol, DBError> {
+    if server.selected_db == 0 {
+        return Ok(Protocol::err("FT commands are not allowed on DB 0"));
+    }
+    // Enforce Tantivy backend for selected DB
+    let is_tantivy = crate::admin_meta::get_database_backend(
+        &server.option.dir,
+        server.option.backend.clone(),
+        &server.option.admin_secret,
+        server.selected_db,
+    )
+    .ok()
+    .flatten()
+    .map(|b| matches!(b, crate::options::BackendType::Tantivy))
+    .unwrap_or(false);
+    if !is_tantivy {
+        return Ok(Protocol::err("ERR DB backend is not Tantivy; FT.* commands are not allowed"));
+    }
+    let indexes = server.search_indexes.read().unwrap();
+    let search_index = indexes
+        .get(&index_name)
+        .ok_or_else(|| DBError(format!("Index '{}' not found", index_name)))?;
+    let info = search_index.get_info()?;
+
+    // Format info as Redis protocol
+    let mut response = Vec::new();
+    response.push(Protocol::BulkString("index_name".to_string()));
+    response.push(Protocol::BulkString(info.name));
+    response.push(Protocol::BulkString("num_docs".to_string()));
+    response.push(Protocol::BulkString(info.num_docs.to_string()));
+    response.push(Protocol::BulkString("num_fields".to_string()));
+    response.push(Protocol::BulkString(info.fields.len().to_string()));
+    response.push(Protocol::BulkString("fields".to_string()));
+    let fields_str = info
+        .fields
+        .iter()
+        .map(|f| format!("{}:{}", f.name, f.field_type))
+        .collect::<Vec<_>>()
+        .join(", ");
+    response.push(Protocol::BulkString(fields_str));
+    Ok(Protocol::Array(response))
+}
+
+pub async fn ft_drop_cmd(server: &Server, index_name: String) -> Result<Protocol, DBError> {
+    if server.selected_db == 0 {
+        return Ok(Protocol::err("FT commands are not allowed on DB 0"));
+    }
+    // Enforce Tantivy backend for selected DB
+    let is_tantivy = crate::admin_meta::get_database_backend(
+        &server.option.dir,
+        server.option.backend.clone(),
+        &server.option.admin_secret,
+        server.selected_db,
+    )
+    .ok()
+    .flatten()
+    .map(|b| matches!(b, crate::options::BackendType::Tantivy))
+    .unwrap_or(false);
+    if !is_tantivy {
+        return Ok(Protocol::err("ERR DB backend is not Tantivy; FT.* commands are not allowed"));
+    }
+
+    // Remove from registry
+    {
+        let mut indexes = server.search_indexes.write().unwrap();
+        indexes.remove(&index_name);
+    }
+
+    // Remove the index files from disk
+    let index_path = server.search_index_path().join(&index_name);
+    if index_path.exists() {
+        std::fs::remove_dir_all(&index_path)
+            .map_err(|e| DBError(format!("Failed to remove index files: {}", e)))?;
+    }
+
+    Ok(Protocol::SimpleString("OK".to_string()))
+}

src/server.rs

@@ -9,12 +9,10 @@ use std::sync::atomic::{AtomicU64, Ordering};
 
 use crate::cmd::Cmd;
 use crate::error::DBError;
-use crate::lance_store::LanceStore;
 use crate::options;
 use crate::protocol::Protocol;
-use crate::storage::Storage;
-use crate::storage_sled::SledStorage;
 use crate::storage_trait::StorageBackend;
+use crate::admin_meta;
 
 #[derive(Clone)]
 pub struct Server {
@@ -23,13 +21,14 @@ pub struct Server {
     pub client_name: Option<String>,
     pub selected_db: u64, // Changed from usize to u64
     pub queued_cmd: Option<Vec<(Cmd, Protocol)>>,
+    pub current_permissions: Option<crate::rpc::Permissions>,
+
+    // In-memory registry of Tantivy search indexes for this server
+    pub search_indexes: Arc<std::sync::RwLock<HashMap<String, Arc<crate::tantivy_search::TantivySearch>>>>,
 
     // BLPOP waiter registry: per (db_index, key) FIFO of waiters
     pub list_waiters: Arc<Mutex<HashMap<u64, HashMap<String, Vec<Waiter>>>>>,
     pub waiter_seq: Arc<AtomicU64>,
-    
-    // Lance vector store
-    pub lance_store: Option<Arc<LanceStore>>,
 }
 
 pub struct Waiter {
@@ -46,81 +45,70 @@ pub enum PopSide {
 
 impl Server {
     pub async fn new(option: options::DBOption) -> Self {
-        // Initialize Lance store
-        let lance_data_dir = std::path::PathBuf::from(&option.dir).join("lance");
-        let lance_store = match LanceStore::new(lance_data_dir).await {
-            Ok(store) => Some(Arc::new(store)),
-            Err(e) => {
-                eprintln!("Warning: Failed to initialize Lance store: {}", e.0);
-                None
-            }
-        };
-
         Server {
             db_cache: Arc::new(std::sync::RwLock::new(HashMap::new())),
             option,
             client_name: None,
             selected_db: 0,
             queued_cmd: None,
-
+            current_permissions: None,
+    
+            search_indexes: Arc::new(std::sync::RwLock::new(HashMap::new())),
             list_waiters: Arc::new(Mutex::new(HashMap::new())),
             waiter_seq: Arc::new(AtomicU64::new(1)),
-            lance_store,
         }
     }
 
-    pub fn lance_store(&self) -> Result<Arc<LanceStore>, DBError> {
-        self.lance_store
-            .as_ref()
-            .cloned()
-            .ok_or_else(|| DBError("Lance store not initialized".to_string()))
+    // Path where search indexes are stored, namespaced per selected DB:
+    // <base_dir>/search_indexes/<db_id>
+    pub fn search_index_path(&self) -> std::path::PathBuf {
+        let base = std::path::PathBuf::from(&self.option.dir)
+            .join("search_indexes")
+            .join(self.selected_db.to_string());
+        if !base.exists() {
+            let _ = std::fs::create_dir_all(&base);
+        }
+        base
     }
 
     pub fn current_storage(&self) -> Result<Arc<dyn StorageBackend>, DBError> {
         let mut cache = self.db_cache.write().unwrap();
-        
+
         if let Some(storage) = cache.get(&self.selected_db) {
             return Ok(storage.clone());
         }
-        
-        
-        // Create new database file
-        let db_file_path = std::path::PathBuf::from(self.option.dir.clone())
-            .join(format!("{}.db", self.selected_db));
-        
-        // Ensure the directory exists before creating the database file
-        if let Some(parent_dir) = db_file_path.parent() {
-            std::fs::create_dir_all(parent_dir).map_err(|e| {
-                DBError(format!("Failed to create directory {}: {}", parent_dir.display(), e))
-            })?;
-        }
-        
-        println!("Creating new db file: {}", db_file_path.display());
-        
-        let storage: Arc<dyn StorageBackend> = match self.option.backend {
-            options::BackendType::Redb => {
-                Arc::new(Storage::new(
-                    db_file_path,
-                    self.should_encrypt_db(self.selected_db),
-                    self.option.encryption_key.as_deref()
-                )?)
-            }
-            options::BackendType::Sled => {
-                Arc::new(SledStorage::new(
-                    db_file_path,
-                    self.should_encrypt_db(self.selected_db),
-                    self.option.encryption_key.as_deref()
-                )?)
-            }
+
+        // Use process-wide shared handles to avoid sled/reDB double-open lock contention.
+        let storage = if self.selected_db == 0 {
+            // Admin DB 0: always via singleton
+            admin_meta::open_admin_storage(
+                &self.option.dir,
+                self.option.backend.clone(),
+                &self.option.admin_secret,
+            )?
+        } else {
+            // Data DBs: via global registry keyed by id
+            admin_meta::open_data_storage(
+                &self.option.dir,
+                self.option.backend.clone(),
+                &self.option.admin_secret,
+                self.selected_db,
+            )?
         };
-        
+
         cache.insert(self.selected_db, storage.clone());
         Ok(storage)
     }
     
-    fn should_encrypt_db(&self, db_index: u64) -> bool {
-        // DB 0-9 are non-encrypted, DB 10+ are encrypted
-        self.option.encrypt && db_index >= 10
+
+    /// Check if current permissions allow read operations
+    pub fn has_read_permission(&self) -> bool {
+        matches!(self.current_permissions, Some(crate::rpc::Permissions::Read) | Some(crate::rpc::Permissions::ReadWrite))
+    }
+
+    /// Check if current permissions allow write operations
+    pub fn has_write_permission(&self) -> bool {
+        matches!(self.current_permissions, Some(crate::rpc::Permissions::ReadWrite))
     }
 
     // ----- BLPOP waiter helpers -----

src/sym.rs

@@ -0,0 +1,123 @@
+//! sym.rs — Stateless symmetric encryption (Phase 1)
+//!
+//! Commands implemented (RESP):
+//! - SYM KEYGEN
+//! - SYM ENCRYPT <key_b64> <message>
+//! - SYM DECRYPT <key_b64> <ciphertext_b64>
+//!
+//! Notes:
+//! - Raw key: exactly 32 bytes, provided as Base64 in commands.
+//! - Cipher: XChaCha20-Poly1305 (AEAD) without AAD in Phase 1
+//! - Ciphertext binary layout: [version:1][nonce:24][ciphertext||tag]
+//! - Encoding for wire I/O: Base64
+
+use base64::{engine::general_purpose::STANDARD as B64, Engine as _};
+use chacha20poly1305::{
+    aead::{Aead, KeyInit, OsRng},
+    XChaCha20Poly1305, XNonce,
+};
+use rand::RngCore;
+
+use crate::protocol::Protocol;
+
+const VERSION: u8 = 1;
+const NONCE_LEN: usize = 24;
+const TAG_LEN: usize = 16;
+
+#[derive(Debug)]
+pub enum SymWireError {
+    InvalidKey,
+    BadEncoding,
+    BadFormat,
+    BadVersion(u8),
+    Crypto,
+}
+
+impl SymWireError {
+    fn to_protocol(self) -> Protocol {
+        match self {
+            SymWireError::InvalidKey => Protocol::err("ERR sym: invalid key"),
+            SymWireError::BadEncoding => Protocol::err("ERR sym: bad encoding"),
+            SymWireError::BadFormat => Protocol::err("ERR sym: bad format"),
+            SymWireError::BadVersion(v) => Protocol::err(&format!("ERR sym: unsupported version {}", v)),
+            SymWireError::Crypto => Protocol::err("ERR sym: auth failed"),
+        }
+    }
+}
+
+fn decode_key_b64(s: &str) -> Result<chacha20poly1305::Key, SymWireError> {
+    let bytes = B64.decode(s.as_bytes()).map_err(|_| SymWireError::BadEncoding)?;
+    if bytes.len() != 32 {
+        return Err(SymWireError::InvalidKey);
+    }
+    Ok(chacha20poly1305::Key::from_slice(&bytes).to_owned())
+}
+
+fn encrypt_blob(key: &chacha20poly1305::Key, plaintext: &[u8]) -> Result<Vec<u8>, SymWireError> {
+    let cipher = XChaCha20Poly1305::new(key);
+
+    let mut nonce_bytes = [0u8; NONCE_LEN];
+    OsRng.fill_bytes(&mut nonce_bytes);
+    let nonce = XNonce::from_slice(&nonce_bytes);
+
+    let mut out = Vec::with_capacity(1 + NONCE_LEN + plaintext.len() + TAG_LEN);
+    out.push(VERSION);
+    out.extend_from_slice(&nonce_bytes);
+
+    let ct = cipher.encrypt(nonce, plaintext).map_err(|_| SymWireError::Crypto)?;
+    out.extend_from_slice(&ct);
+    Ok(out)
+}
+
+fn decrypt_blob(key: &chacha20poly1305::Key, blob: &[u8]) -> Result<Vec<u8>, SymWireError> {
+    if blob.len() < 1 + NONCE_LEN + TAG_LEN {
+        return Err(SymWireError::BadFormat);
+    }
+    let ver = blob[0];
+    if ver != VERSION {
+        return Err(SymWireError::BadVersion(ver));
+    }
+    let nonce = XNonce::from_slice(&blob[1..1 + NONCE_LEN]);
+    let ct = &blob[1 + NONCE_LEN..];
+
+    let cipher = XChaCha20Poly1305::new(key);
+    cipher.decrypt(nonce, ct).map_err(|_| SymWireError::Crypto)
+}
+
+// ---------- Command handlers (RESP) ----------
+
+pub async fn cmd_sym_keygen() -> Protocol {
+    let mut key_bytes = [0u8; 32];
+    OsRng.fill_bytes(&mut key_bytes);
+    let key_b64 = B64.encode(key_bytes);
+    Protocol::BulkString(key_b64)
+}
+
+pub async fn cmd_sym_encrypt(key_b64: &str, message: &str) -> Protocol {
+    let key = match decode_key_b64(key_b64) {
+        Ok(k) => k,
+        Err(e) => return e.to_protocol(),
+    };
+    match encrypt_blob(&key, message.as_bytes()) {
+        Ok(blob) => Protocol::BulkString(B64.encode(blob)),
+        Err(e) => e.to_protocol(),
+    }
+}
+
+pub async fn cmd_sym_decrypt(key_b64: &str, ct_b64: &str) -> Protocol {
+    let key = match decode_key_b64(key_b64) {
+        Ok(k) => k,
+        Err(e) => return e.to_protocol(),
+    };
+    let blob = match B64.decode(ct_b64.as_bytes()) {
+        Ok(b) => b,
+        Err(_) => return SymWireError::BadEncoding.to_protocol(),
+    };
+    match decrypt_blob(&key, &blob) {
+        Ok(pt) => match String::from_utf8(pt) {
+            Ok(s) => Protocol::BulkString(s),
+            Err(_) => Protocol::err("ERR sym: invalid UTF-8 plaintext"),
+        },
+        Err(e) => e.to_protocol(),
+    }
+}

src/tantivy_search.rs

@@ -0,0 +1,667 @@
+use crate::error::DBError;
+use serde::{Deserialize, Serialize};
+use std::collections::HashMap;
+use std::path::PathBuf;
+use std::sync::{Arc, RwLock};
+use tantivy::{
+    collector::TopDocs,
+    directory::MmapDirectory,
+    query::{BooleanQuery, Occur, Query, QueryParser, TermQuery},
+    schema::{
+        DateOptions, Field, IndexRecordOption, NumericOptions, Schema, TextFieldIndexing, TextOptions, STORED, STRING,
+    },
+    tokenizer::TokenizerManager,
+    DateTime, Index, IndexReader, IndexWriter, TantivyDocument, Term,
+};
+use tantivy::schema::Value;
+
+#[derive(Debug, Clone, Serialize, Deserialize)]
+pub enum FieldDef {
+    Text {
+        stored: bool,
+        indexed: bool,
+        tokenized: bool,
+        fast: bool,
+    },
+    Numeric {
+        stored: bool,
+        indexed: bool,
+        fast: bool,
+        precision: NumericType,
+    },
+    Tag {
+        stored: bool,
+        separator: String,
+        case_sensitive: bool,
+    },
+    Geo {
+        stored: bool,
+    },
+}
+
+#[derive(Debug, Clone, Serialize, Deserialize)]
+pub enum NumericType {
+    I64,
+    U64,
+    F64,
+    Date,
+}
+
+pub struct IndexSchema {
+    schema: Schema,
+    fields: HashMap<String, (Field, FieldDef)>,
+    default_search_fields: Vec<Field>,
+}
+
+pub struct TantivySearch {
+    index: Index,
+    writer: Arc<RwLock<IndexWriter>>,
+    reader: IndexReader,
+    index_schema: IndexSchema,
+    name: String,
+    config: IndexConfig,
+}
+
+#[derive(Debug, Clone, Serialize, Deserialize)]
+pub struct IndexConfig {
+    pub language: String,
+    pub stopwords: Vec<String>,
+    pub stemming: bool,
+    pub max_doc_count: Option<usize>,
+    pub default_score: f64,
+}
+
+impl Default for IndexConfig {
+    fn default() -> Self {
+        IndexConfig {
+            language: "english".to_string(),
+            stopwords: vec![],
+            stemming: true,
+            max_doc_count: None,
+            default_score: 1.0,
+        }
+    }
+}
+
+impl TantivySearch {
+    pub fn new_with_schema(
+        base_path: PathBuf,
+        name: String,
+        field_definitions: Vec<(String, FieldDef)>,
+        config: Option<IndexConfig>,
+    ) -> Result<Self, DBError> {
+        let index_path = base_path.join(&name);
+        std::fs::create_dir_all(&index_path)
+            .map_err(|e| DBError(format!("Failed to create index dir: {}", e)))?;
+
+        // Build schema from field definitions
+        let mut schema_builder = Schema::builder();
+        let mut fields = HashMap::new();
+        let mut default_search_fields = Vec::new();
+
+        // Always add a document ID field
+        let id_field = schema_builder.add_text_field("_id", STRING | STORED);
+        fields.insert(
+            "_id".to_string(),
+            (
+                id_field,
+                FieldDef::Text {
+                    stored: true,
+                    indexed: true,
+                    tokenized: false,
+                    fast: false,
+                },
+            ),
+        );
+
+        // Add user-defined fields
+        for (field_name, field_def) in field_definitions {
+            let field = match &field_def {
+                FieldDef::Text {
+                    stored,
+                    indexed,
+                    tokenized,
+                    fast: _fast,
+                } => {
+                    let mut text_options = TextOptions::default();
+                    if *stored {
+                        text_options = text_options.set_stored();
+                    }
+                    if *indexed {
+                        let indexing_options = if *tokenized {
+                            TextFieldIndexing::default()
+                                .set_tokenizer("default")
+                                .set_index_option(IndexRecordOption::WithFreqsAndPositions)
+                        } else {
+                            TextFieldIndexing::default()
+                                .set_tokenizer("raw")
+                                .set_index_option(IndexRecordOption::Basic)
+                        };
+                        text_options = text_options.set_indexing_options(indexing_options);
+                        let f = schema_builder.add_text_field(&field_name, text_options);
+                        if *tokenized {
+                            default_search_fields.push(f);
+                        }
+                        f
+                    } else {
+                        schema_builder.add_text_field(&field_name, text_options)
+                    }
+                }
+                FieldDef::Numeric {
+                    stored,
+                    indexed,
+                    fast,
+                    precision,
+                } => match precision {
+                    NumericType::I64 => {
+                        let mut opts = NumericOptions::default();
+                        if *stored {
+                            opts = opts.set_stored();
+                        }
+                        if *indexed {
+                            opts = opts.set_indexed();
+                        }
+                        if *fast {
+                            opts = opts.set_fast();
+                        }
+                        schema_builder.add_i64_field(&field_name, opts)
+                    }
+                    NumericType::U64 => {
+                        let mut opts = NumericOptions::default();
+                        if *stored {
+                            opts = opts.set_stored();
+                        }
+                        if *indexed {
+                            opts = opts.set_indexed();
+                        }
+                        if *fast {
+                            opts = opts.set_fast();
+                        }
+                        schema_builder.add_u64_field(&field_name, opts)
+                    }
+                    NumericType::F64 => {
+                        let mut opts = NumericOptions::default();
+                        if *stored {
+                            opts = opts.set_stored();
+                        }
+                        if *indexed {
+                            opts = opts.set_indexed();
+                        }
+                        if *fast {
+                            opts = opts.set_fast();
+                        }
+                        schema_builder.add_f64_field(&field_name, opts)
+                    }
+                    NumericType::Date => {
+                        let mut opts = DateOptions::default();
+                        if *stored {
+                            opts = opts.set_stored();
+                        }
+                        if *indexed {
+                            opts = opts.set_indexed();
+                        }
+                        if *fast {
+                            opts = opts.set_fast();
+                        }
+                        schema_builder.add_date_field(&field_name, opts)
+                    }
+                },
+                FieldDef::Tag {
+                    stored,
+                    separator: _,
+                    case_sensitive: _,
+                } => {
+                    let mut text_options = TextOptions::default();
+                    if *stored {
+                        text_options = text_options.set_stored();
+                    }
+                    text_options = text_options.set_indexing_options(
+                        TextFieldIndexing::default()
+                            .set_tokenizer("raw")
+                            .set_index_option(IndexRecordOption::Basic),
+                    );
+                    schema_builder.add_text_field(&field_name, text_options)
+                }
+                FieldDef::Geo { stored } => {
+                    // For now, store as two f64 fields for lat/lon
+                    let mut opts = NumericOptions::default();
+                    if *stored {
+                        opts = opts.set_stored();
+                    }
+                    opts = opts.set_indexed().set_fast();
+                    let lat_field =
+                        schema_builder.add_f64_field(&format!("{}_lat", field_name), opts.clone());
+                    let lon_field =
+                        schema_builder.add_f64_field(&format!("{}_lon", field_name), opts);
+                    fields.insert(
+                        format!("{}_lat", field_name),
+                        (
+                            lat_field,
+                            FieldDef::Numeric {
+                                stored: *stored,
+                                indexed: true,
+                                fast: true,
+                                precision: NumericType::F64,
+                            },
+                        ),
+                    );
+                    fields.insert(
+                        format!("{}_lon", field_name),
+                        (
+                            lon_field,
+                            FieldDef::Numeric {
+                                stored: *stored,
+                                indexed: true,
+                                fast: true,
+                                precision: NumericType::F64,
+                            },
+                        ),
+                    );
+                    continue; // Skip adding the geo field itself
+                }
+            };
+            fields.insert(field_name.clone(), (field, field_def));
+        }
+
+        let schema = schema_builder.build();
+        let index_schema = IndexSchema {
+            schema: schema.clone(),
+            fields,
+            default_search_fields,
+        };
+
+        // Create or open index
+        let dir = MmapDirectory::open(&index_path)
+            .map_err(|e| DBError(format!("Failed to open index directory: {}", e)))?;
+        let mut index =
+            Index::open_or_create(dir, schema).map_err(|e| DBError(format!("Failed to create index: {}", e)))?;
+
+        // Configure tokenizers
+        let tokenizer_manager = TokenizerManager::default();
+        index.set_tokenizers(tokenizer_manager);
+
+        let writer = index
+            .writer(15_000_000)
+            .map_err(|e| DBError(format!("Failed to create index writer: {}", e)))?;
+        let reader = index
+            .reader()
+            .map_err(|e| DBError(format!("Failed to create reader: {}", e)))?;
+
+        let config = config.unwrap_or_default();
+
+        Ok(TantivySearch {
+            index,
+            writer: Arc::new(RwLock::new(writer)),
+            reader,
+            index_schema,
+            name,
+            config,
+        })
+    }
+
+    pub fn add_document_with_fields(
+        &self,
+        doc_id: &str,
+        fields: HashMap<String, String>,
+    ) -> Result<(), DBError> {
+        let mut writer = self
+            .writer
+            .write()
+            .map_err(|e| DBError(format!("Failed to acquire writer lock: {}", e)))?;
+
+        // Delete existing document with same ID
+        if let Some((id_field, _)) = self.index_schema.fields.get("_id") {
+            writer.delete_term(Term::from_field_text(*id_field, doc_id));
+        }
+
+        // Create new document
+        let mut doc = tantivy::doc!();
+
+        // Add document ID
+        if let Some((id_field, _)) = self.index_schema.fields.get("_id") {
+            doc.add_text(*id_field, doc_id);
+        }
+
+        // Add other fields based on schema
+        for (field_name, field_value) in fields {
+            if let Some((field, field_def)) = self.index_schema.fields.get(&field_name) {
+                match field_def {
+                    FieldDef::Text { .. } => {
+                        doc.add_text(*field, &field_value);
+                    }
+                    FieldDef::Numeric { precision, .. } => match precision {
+                        NumericType::I64 => {
+                            if let Ok(v) = field_value.parse::<i64>() {
+                                doc.add_i64(*field, v);
+                            }
+                        }
+                        NumericType::U64 => {
+                            if let Ok(v) = field_value.parse::<u64>() {
+                                doc.add_u64(*field, v);
+                            }
+                        }
+                        NumericType::F64 => {
+                            if let Ok(v) = field_value.parse::<f64>() {
+                                doc.add_f64(*field, v);
+                            }
+                        }
+                        NumericType::Date => {
+                            if let Ok(v) = field_value.parse::<i64>() {
+                                doc.add_date(*field, DateTime::from_timestamp_millis(v));
+                            }
+                        }
+                    },
+                    FieldDef::Tag {
+                        separator,
+                        case_sensitive,
+                        ..
+                    } => {
+                        let tags = if !case_sensitive {
+                            field_value.to_lowercase()
+                        } else {
+                            field_value.clone()
+                        };
+                        for tag in tags.split(separator.as_str()) {
+                            doc.add_text(*field, tag.trim());
+                        }
+                    }
+                    FieldDef::Geo { .. } => {
+                        let parts: Vec<&str> = field_value.split(',').collect();
+                        if parts.len() == 2 {
+                            if let (Ok(lat), Ok(lon)) =
+                                (parts[0].parse::<f64>(), parts[1].parse::<f64>())
+                            {
+                                if let Some((lat_field, _)) =
+                                    self.index_schema.fields.get(&format!("{}_lat", field_name))
+                                {
+                                    doc.add_f64(*lat_field, lat);
+                                }
+                                if let Some((lon_field, _)) =
+                                    self.index_schema.fields.get(&format!("{}_lon", field_name))
+                                {
+                                    doc.add_f64(*lon_field, lon);
+                                }
+                            }
+                        }
+                    }
+                }
+            }
+        }
+
+        writer
+            .add_document(doc)
+            .map_err(|e| DBError(format!("Failed to add document: {}", e)))?;
+        writer
+            .commit()
+            .map_err(|e| DBError(format!("Failed to commit: {}", e)))?;
+        Ok(())
+    }
+
+    pub fn search_with_options(
+        &self,
+        query_str: &str,
+        options: SearchOptions,
+    ) -> Result<SearchResults, DBError> {
+        let searcher = self.reader.searcher();
+
+        // Ensure we have searchable fields
+        if self.index_schema.default_search_fields.is_empty() {
+            return Err(DBError("No searchable fields defined in schema".to_string()));
+        }
+
+        // Parse query based on search fields
+        let query_parser = QueryParser::for_index(
+            &self.index,
+            self.index_schema.default_search_fields.clone(),
+        );
+        let parsed_query = query_parser
+            .parse_query(query_str)
+            .map_err(|e| DBError(format!("Failed to parse query: {}", e)))?;
+        let mut clauses: Vec<(Occur, Box<dyn Query>)> = vec![(Occur::Must, parsed_query)];
+
+        // Apply filters if any
+        for filter in options.filters {
+            if let Some((field, field_def)) = self.index_schema.fields.get(&filter.field) {
+                match filter.filter_type {
+                    FilterType::Equals(value) => {
+                        match field_def {
+                            FieldDef::Text { .. } | FieldDef::Tag { .. } => {
+                                let term_query =
+                                    TermQuery::new(Term::from_field_text(*field, &value), IndexRecordOption::Basic);
+                                clauses.push((Occur::Must, Box::new(term_query)));
+                            }
+                            FieldDef::Numeric { precision, .. } => {
+                                // Equals on numeric fields: parse to the right numeric type and use term query
+                                match precision {
+                                    NumericType::I64 => {
+                                        if let Ok(v) = value.parse::<i64>() {
+                                            let term = Term::from_field_i64(*field, v);
+                                            let tq = TermQuery::new(term, IndexRecordOption::Basic);
+                                            clauses.push((Occur::Must, Box::new(tq)));
+                                        }
+                                    }
+                                    NumericType::U64 => {
+                                        if let Ok(v) = value.parse::<u64>() {
+                                            let term = Term::from_field_u64(*field, v);
+                                            let tq = TermQuery::new(term, IndexRecordOption::Basic);
+                                            clauses.push((Occur::Must, Box::new(tq)));
+                                        }
+                                    }
+                                    NumericType::F64 => {
+                                        if let Ok(v) = value.parse::<f64>() {
+                                            let term = Term::from_field_f64(*field, v);
+                                            let tq = TermQuery::new(term, IndexRecordOption::Basic);
+                                            clauses.push((Occur::Must, Box::new(tq)));
+                                        }
+                                    }
+                                    NumericType::Date => {
+                                        if let Ok(v) = value.parse::<i64>() {
+                                            let dt = DateTime::from_timestamp_millis(v);
+                                            let term = Term::from_field_date(*field, dt);
+                                            let tq = TermQuery::new(term, IndexRecordOption::Basic);
+                                            clauses.push((Occur::Must, Box::new(tq)));
+                                        }
+                                    }
+                                }
+                            }
+                            FieldDef::Geo { .. } => {
+                                // Geo equals isn't supported in this simplified version
+                            }
+                        }
+                    }
+                    FilterType::Range { .. } => {
+                        // TODO: Implement numeric range queries by building a RangeQuery per type
+                    }
+                    FilterType::InSet(values) => {
+                        // OR across values
+                        let mut sub_clauses: Vec<(Occur, Box<dyn Query>)> = vec![];
+                        for value in values {
+                            let term_query = TermQuery::new(
+                                Term::from_field_text(*field, &value),
+                                IndexRecordOption::Basic,
+                            );
+                            sub_clauses.push((Occur::Should, Box::new(term_query)));
+                        }
+                        clauses.push((Occur::Must, Box::new(BooleanQuery::new(sub_clauses))));
+                    }
+                }
+            }
+        }
+
+        let final_query: Box<dyn Query> = if clauses.len() == 1 {
+            clauses.pop().unwrap().1
+        } else {
+            Box::new(BooleanQuery::new(clauses))
+        };
+
+        // Execute search
+        let top_docs = searcher
+            .search(&*final_query, &TopDocs::with_limit(options.limit + options.offset))
+            .map_err(|e| DBError(format!("Search failed: {}", e)))?;
+        let total_hits = top_docs.len();
+        let mut documents = Vec::new();
+
+        for (score, doc_address) in top_docs.into_iter().skip(options.offset).take(options.limit) {
+            let retrieved_doc: TantivyDocument = searcher
+                .doc(doc_address)
+                .map_err(|e| DBError(format!("Failed to retrieve doc: {}", e)))?;
+
+            let mut doc_fields = HashMap::new();
+
+            // Extract stored fields (or synthesize)
+            for (field_name, (field, field_def)) in &self.index_schema.fields {
+                match field_def {
+                    FieldDef::Text { stored, .. } | FieldDef::Tag { stored, .. } => {
+                        if *stored {
+                            if let Some(value) = retrieved_doc.get_first(*field) {
+                                if let Some(text) = value.as_str() {
+                                    doc_fields.insert(field_name.clone(), text.to_string());
+                                }
+                            }
+                        }
+                    }
+                    FieldDef::Numeric {
+                        stored, precision, ..
+                    } => {
+                        if *stored {
+                            let value_str = match precision {
+                                NumericType::I64 => retrieved_doc
+                                    .get_first(*field)
+                                    .and_then(|v| v.as_i64())
+                                    .map(|v| v.to_string()),
+                                NumericType::U64 => retrieved_doc
+                                    .get_first(*field)
+                                    .and_then(|v| v.as_u64())
+                                    .map(|v| v.to_string()),
+                                NumericType::F64 => retrieved_doc
+                                    .get_first(*field)
+                                    .and_then(|v| v.as_f64())
+                                    .map(|v| v.to_string()),
+                                NumericType::Date => retrieved_doc
+                                    .get_first(*field)
+                                    .and_then(|v| v.as_datetime())
+                                    .map(|v| v.into_timestamp_millis().to_string()),
+                            };
+                            if let Some(v) = value_str {
+                                doc_fields.insert(field_name.clone(), v);
+                            }
+                        }
+                    }
+                    FieldDef::Geo { stored } => {
+                        if *stored {
+                            let lat_field = self
+                                .index_schema
+                                .fields
+                                .get(&format!("{}_lat", field_name))
+                                .unwrap()
+                                .0;
+                            let lon_field = self
+                                .index_schema
+                                .fields
+                                .get(&format!("{}_lon", field_name))
+                                .unwrap()
+                                .0;
+                            let lat = retrieved_doc.get_first(lat_field).and_then(|v| v.as_f64());
+                            let lon = retrieved_doc.get_first(lon_field).and_then(|v| v.as_f64());
+                            if let (Some(lat), Some(lon)) = (lat, lon) {
+                                doc_fields.insert(field_name.clone(), format!("{},{}", lat, lon));
+                            }
+                        }
+                    }
+                }
+            }
+
+            documents.push(SearchDocument {
+                fields: doc_fields,
+                score,
+            });
+        }
+
+        Ok(SearchResults {
+            total: total_hits,
+            documents,
+        })
+    }
+
+    pub fn get_info(&self) -> Result<IndexInfo, DBError> {
+        let searcher = self.reader.searcher();
+        let num_docs = searcher.num_docs();
+        let fields_info: Vec<FieldInfo> = self
+            .index_schema
+            .fields
+            .iter()
+            .map(|(name, (_, def))| FieldInfo {
+                name: name.clone(),
+                field_type: format!("{:?}", def),
+            })
+            .collect();
+        Ok(IndexInfo {
+            name: self.name.clone(),
+            num_docs,
+            fields: fields_info,
+            config: self.config.clone(),
+        })
+    }
+}
+
+#[derive(Debug, Clone)]
+pub struct SearchOptions {
+    pub limit: usize,
+    pub offset: usize,
+    pub filters: Vec<Filter>,
+    pub sort_by: Option<String>,
+    pub return_fields: Option<Vec<String>>,
+    pub highlight: bool,
+}
+
+impl Default for SearchOptions {
+    fn default() -> Self {
+        SearchOptions {
+            limit: 10,
+            offset: 0,
+            filters: vec![],
+            sort_by: None,
+            return_fields: None,
+            highlight: false,
+        }
+    }
+}
+
+#[derive(Debug, Clone)]
+pub struct Filter {
+    pub field: String,
+    pub filter_type: FilterType,
+}
+
+#[derive(Debug, Clone)]
+pub enum FilterType {
+    Equals(String),
+    Range { min: String, max: String },
+    InSet(Vec<String>),
+}
+
+#[derive(Debug)]
+pub struct SearchResults {
+    pub total: usize,
+    pub documents: Vec<SearchDocument>,
+}
+
+#[derive(Debug)]
+pub struct SearchDocument {
+    pub fields: HashMap<String, String>,
+    pub score: f32,
+}
+
+#[derive(Debug, Serialize, Deserialize)]
+pub struct IndexInfo {
+    pub name: String,
+    pub num_docs: u64,
+    pub fields: Vec<FieldInfo>,
+    pub config: IndexConfig,
+}
+
+#[derive(Debug, Serialize, Deserialize)]
+pub struct FieldInfo {
+    pub name: String,
+    pub field_type: String,
+}

tests/debug_hset.rs

@@ -28,6 +28,7 @@ async fn debug_hset_simple() {
         encrypt: false,
         encryption_key: None,
         backend: herodb::options::BackendType::Redb,
+        admin_secret: "test-admin".to_string(),
     };
     
     let mut server = Server::new(option).await;
@@ -48,6 +49,12 @@ async fn debug_hset_simple() {
     sleep(Duration::from_millis(200)).await;
     
     let mut stream = TcpStream::connect(format!("127.0.0.1:{}", port)).await.unwrap();
+    // Acquire ReadWrite permissions on this connection
+    let resp = send_command(
+        &mut stream,
+        "*4\r\n$6\r\nSELECT\r\n$1\r\n0\r\n$3\r\nKEY\r\n$10\r\ntest-admin\r\n",
+    ).await;
+    assert!(resp.contains("OK"), "Failed SELECT handshake: {}", resp);
     
     // Test simple HSET
     println!("Testing HSET...");

tests/debug_hset_simple.rs

@@ -19,6 +19,7 @@ async fn debug_hset_return_value() {
         encrypt: false,
         encryption_key: None,
         backend: herodb::options::BackendType::Redb,
+        admin_secret: "test-admin".to_string(),
     };
     
     let mut server = Server::new(option).await;
@@ -40,12 +41,19 @@ async fn debug_hset_return_value() {
     
     // Connect and test HSET
     let mut stream = TcpStream::connect("127.0.0.1:16390").await.unwrap();
+
+    // Acquire ReadWrite permissions for this new connection
+    let handshake = "*4\r\n$6\r\nSELECT\r\n$1\r\n0\r\n$3\r\nKEY\r\n$10\r\ntest-admin\r\n";
+    stream.write_all(handshake.as_bytes()).await.unwrap();
+    let mut buffer = [0; 1024];
+    let n = stream.read(&mut buffer).await.unwrap();
+    let resp = String::from_utf8_lossy(&buffer[..n]);
+    assert!(resp.contains("OK"), "Failed SELECT handshake: {}", resp);
     
     // Send HSET command
     let cmd = "*4\r\n$4\r\nHSET\r\n$4\r\nhash\r\n$6\r\nfield1\r\n$6\r\nvalue1\r\n";
     stream.write_all(cmd.as_bytes()).await.unwrap();
     
-    let mut buffer = [0; 1024];
     let n = stream.read(&mut buffer).await.unwrap();
     let response = String::from_utf8_lossy(&buffer[..n]);
     

tests/redis_integration_tests.rs

@@ -12,7 +12,15 @@ fn get_redis_connection(port: u16) -> Connection {
         match client.get_connection() {
             Ok(mut conn) => {
                 if redis::cmd("PING").query::<String>(&mut conn).is_ok() {
-                    return conn;
+                    // Acquire ReadWrite permissions on this connection
+                    let sel: RedisResult<String> = redis::cmd("SELECT")
+                        .arg(0)
+                        .arg("KEY")
+                        .arg("test-admin")
+                        .query(&mut conn);
+                    if sel.is_ok() {
+                        return conn;
+                    }
                 }
             }
             Err(e) => {
@@ -78,6 +86,8 @@ fn setup_server() -> (ServerProcessGuard, u16) {
             "--port",
             &port.to_string(),
             "--debug",
+            "--admin-secret",
+            "test-admin",
         ])
         .spawn()
         .expect("Failed to start server process");

tests/redis_tests.rs

@@ -23,18 +23,29 @@ async fn start_test_server(test_name: &str) -> (Server, u16) {
         encrypt: false,
         encryption_key: None,
         backend: herodb::options::BackendType::Redb,
+        admin_secret: "test-admin".to_string(),
     };
     
     let server = Server::new(option).await;
     (server, port)
 }
 
-// Helper function to connect to the test server
+ // Helper function to connect to the test server
 async fn connect_to_server(port: u16) -> TcpStream {
     let mut attempts = 0;
     loop {
         match TcpStream::connect(format!("127.0.0.1:{}", port)).await {
-            Ok(stream) => return stream,
+            Ok(mut stream) => {
+                // Obtain ReadWrite permissions for this connection by selecting DB 0 with admin key
+                let resp = send_command(
+                    &mut stream,
+                    "*4\r\n$6\r\nSELECT\r\n$1\r\n0\r\n$3\r\nKEY\r\n$10\r\ntest-admin\r\n",
+                ).await;
+                if !resp.contains("OK") {
+                    panic!("Failed to acquire write permissions via SELECT 0 KEY test-admin: {}", resp);
+                }
+                return stream;
+            }
             Err(_) if attempts < 10 => {
                 attempts += 1;
                 sleep(Duration::from_millis(100)).await;

tests/rpc_tests.rs

@@ -0,0 +1,85 @@
+use herodb::rpc::{BackendType, DatabaseConfig};
+use herodb::admin_meta;
+use herodb::options::BackendType as OptionsBackendType;
+
+#[tokio::test]
+async fn test_rpc_server_basic() {
+    // This test would require starting the RPC server in a separate thread
+    // For now, we'll just test that the types compile correctly
+
+    // Test serialization of types
+    let backend = BackendType::Redb;
+    let config = DatabaseConfig {
+        name: Some("test_db".to_string()),
+        storage_path: Some("/tmp/test".to_string()),
+        max_size: Some(1024 * 1024),
+        redis_version: Some("7.0".to_string()),
+    };
+
+    let backend_json = serde_json::to_string(&backend).unwrap();
+    let config_json = serde_json::to_string(&config).unwrap();
+
+    assert_eq!(backend_json, "\"Redb\"");
+    assert!(config_json.contains("test_db"));
+}
+
+#[tokio::test]
+async fn test_database_config_serialization() {
+    let config = DatabaseConfig {
+        name: Some("my_db".to_string()),
+        storage_path: None,
+        max_size: Some(1000000),
+        redis_version: Some("7.0".to_string()),
+    };
+
+    let json = serde_json::to_value(&config).unwrap();
+    assert_eq!(json["name"], "my_db");
+    assert_eq!(json["max_size"], 1000000);
+    assert_eq!(json["redis_version"], "7.0");
+}
+
+#[tokio::test]
+async fn test_backend_type_serialization() {
+    // Test that both Redb and Sled backends serialize correctly
+    let redb_backend = BackendType::Redb;
+    let sled_backend = BackendType::Sled;
+
+    let redb_json = serde_json::to_string(&redb_backend).unwrap();
+    let sled_json = serde_json::to_string(&sled_backend).unwrap();
+
+    assert_eq!(redb_json, "\"Redb\"");
+    assert_eq!(sled_json, "\"Sled\"");
+
+    // Test deserialization
+    let redb_deserialized: BackendType = serde_json::from_str(&redb_json).unwrap();
+    let sled_deserialized: BackendType = serde_json::from_str(&sled_json).unwrap();
+
+    assert!(matches!(redb_deserialized, BackendType::Redb));
+    assert!(matches!(sled_deserialized, BackendType::Sled));
+}
+
+#[tokio::test]
+async fn test_database_name_persistence() {
+    let base_dir = "/tmp/test_db_name_persistence";
+    let admin_secret = "test-admin-secret";
+    let backend = OptionsBackendType::Redb;
+    let db_id = 1;
+    let test_name = "test-database-name";
+
+    // Clean up any existing test data
+    let _ = std::fs::remove_dir_all(base_dir);
+
+    // Set the database name
+    admin_meta::set_database_name(base_dir, backend.clone(), admin_secret, db_id, test_name)
+        .expect("Failed to set database name");
+
+    // Retrieve the database name
+    let retrieved_name = admin_meta::get_database_name(base_dir, backend, admin_secret, db_id)
+        .expect("Failed to get database name");
+
+    // Verify the name matches
+    assert_eq!(retrieved_name, Some(test_name.to_string()));
+
+    // Clean up
+    let _ = std::fs::remove_dir_all(base_dir);
+}

tests/simple_integration_test.rs

@@ -25,6 +25,7 @@ async fn start_test_server(test_name: &str) -> (Server, u16) {
         encrypt: false,
         encryption_key: None,
         backend: herodb::options::BackendType::Redb,
+        admin_secret: "test-admin".to_string(),
     };
     
     let server = Server::new(option).await;
@@ -34,9 +35,16 @@ async fn start_test_server(test_name: &str) -> (Server, u16) {
 // Helper function to send Redis command and get response
 async fn send_redis_command(port: u16, command: &str) -> String {
     let mut stream = TcpStream::connect(format!("127.0.0.1:{}", port)).await.unwrap();
+    
+    // Acquire ReadWrite permissions on this new connection
+    let handshake = "*4\r\n$6\r\nSELECT\r\n$1\r\n0\r\n$3\r\nKEY\r\n$10\r\ntest-admin\r\n";
+    stream.write_all(handshake.as_bytes()).await.unwrap();
+    let mut buffer = [0; 1024];
+    let _ = stream.read(&mut buffer).await.unwrap(); // Read and ignore the OK for handshake
+    
+    // Now send the intended command
     stream.write_all(command.as_bytes()).await.unwrap();
     
-    let mut buffer = [0; 1024];
     let n = stream.read(&mut buffer).await.unwrap();
     String::from_utf8_lossy(&buffer[..n]).to_string()
 }
@@ -184,12 +192,19 @@ async fn test_transaction_operations() {
     
     sleep(Duration::from_millis(100)).await;
     
-    // Use a single connection for the transaction
+     // Use a single connection for the transaction
     let mut stream = TcpStream::connect(format!("127.0.0.1:{}", port)).await.unwrap();
     
+    // Acquire write permissions for this connection
+    let handshake = "*4\r\n$6\r\nSELECT\r\n$1\r\n0\r\n$3\r\nKEY\r\n$10\r\ntest-admin\r\n";
+    stream.write_all(handshake.as_bytes()).await.unwrap();
+    let mut buffer = [0; 1024];
+    let n = stream.read(&mut buffer).await.unwrap();
+    let resp = String::from_utf8_lossy(&buffer[..n]);
+    assert!(resp.contains("OK"));
+    
     // Test MULTI
     stream.write_all("*1\r\n$5\r\nMULTI\r\n".as_bytes()).await.unwrap();
-    let mut buffer = [0; 1024];
     let n = stream.read(&mut buffer).await.unwrap();
     let response = String::from_utf8_lossy(&buffer[..n]);
     assert!(response.contains("OK"));

tests/simple_redis_test.rs

@@ -23,6 +23,7 @@ async fn start_test_server(test_name: &str) -> (Server, u16) {
         encrypt: false,
         encryption_key: None,
         backend: herodb::options::BackendType::Redb,
+        admin_secret: "test-admin".to_string(),
     };
     
     let server = Server::new(option).await;
@@ -38,12 +39,22 @@ async fn send_command(stream: &mut TcpStream, command: &str) -> String {
     String::from_utf8_lossy(&buffer[..n]).to_string()
 }
 
-// Helper function to connect to the test server
+ // Helper function to connect to the test server
 async fn connect_to_server(port: u16) -> TcpStream {
     let mut attempts = 0;
     loop {
         match TcpStream::connect(format!("127.0.0.1:{}", port)).await {
-            Ok(stream) => return stream,
+            Ok(mut stream) => {
+                // Acquire ReadWrite permissions for this connection
+                let resp = send_command(
+                    &mut stream,
+                    "*4\r\n$6\r\nSELECT\r\n$1\r\n0\r\n$3\r\nKEY\r\n$10\r\ntest-admin\r\n",
+                ).await;
+                if !resp.contains("OK") {
+                    panic!("Failed to acquire write permissions via SELECT 0 KEY test-admin: {}", resp);
+                }
+                return stream;
+            }
             Err(_) if attempts < 10 => {
                 attempts += 1;
                 sleep(Duration::from_millis(100)).await;
@@ -97,14 +108,21 @@ async fn test_hset_clean_db() {
     sleep(Duration::from_millis(200)).await;
     
     let mut stream = connect_to_server(port).await;
-    
-    // Test HSET - should return 1 for new field
-    let response = send_command(&mut stream, "*4\r\n$4\r\nHSET\r\n$4\r\nhash\r\n$6\r\nfield1\r\n$6\r\nvalue1\r\n").await;
+
+    // Ensure clean DB state (admin DB 0 may be shared due to global singleton)
+    let flush = send_command(&mut stream, "*1\r\n$7\r\nFLUSHDB\r\n").await;
+    assert!(flush.contains("OK"), "Failed to FLUSHDB: {}", flush);
+
+    // Test HSET - should return 1 for new field (use a unique key name to avoid collisions)
+    let key = "hash_clean";
+    let hset_cmd = format!("*4\r\n$4\r\nHSET\r\n${}\r\n{}\r\n$6\r\nfield1\r\n$6\r\nvalue1\r\n", key.len(), key);
+    let response = send_command(&mut stream, &hset_cmd).await;
     println!("HSET response: {}", response);
     assert!(response.contains("1"), "Expected HSET to return 1, got: {}", response);
     
     // Test HGET
-    let response = send_command(&mut stream, "*3\r\n$4\r\nHGET\r\n$4\r\nhash\r\n$6\r\nfield1\r\n").await;
+    let hget_cmd = format!("*3\r\n$4\r\nHGET\r\n${}\r\n{}\r\n$6\r\nfield1\r\n", key.len(), key);
+    let response = send_command(&mut stream, &hget_cmd).await;
     println!("HGET response: {}", response);
     assert!(response.contains("value1"));
 }

tests/usage_suite.rs

@@ -23,6 +23,7 @@ async fn start_test_server(test_name: &str) -> (Server, u16) {
         encrypt: false,
         encryption_key: None,
         backend: herodb::options::BackendType::Redb,
+        admin_secret: "test-admin".to_string(),
     };
 
     let server = Server::new(option).await;
@@ -61,7 +62,17 @@ async fn connect(port: u16) -> TcpStream {
     let mut attempts = 0;
     loop {
         match TcpStream::connect(format!("127.0.0.1:{}", port)).await {
-            Ok(s) => return s,
+            Ok(mut s) => {
+                // Acquire ReadWrite permissions for this connection using admin DB 0
+                let resp = send_cmd(&mut s, &["SELECT", "0", "KEY", "test-admin"]).await;
+                assert_contains(&resp, "OK", "SELECT 0 KEY test-admin handshake");
+
+                // Ensure clean slate per test on DB 0
+                let fl = send_cmd(&mut s, &["FLUSHDB"]).await;
+                assert_contains(&fl, "OK", "FLUSHDB after handshake");
+
+                return s;
+            }
             Err(_) if attempts < 30 => {
                 attempts += 1;
                 sleep(Duration::from_millis(100)).await;
@@ -246,9 +257,9 @@ async fn test_01_connection_and_info() {
     let getname = send_cmd(&mut s, &["CLIENT", "GETNAME"]).await;
     assert_contains(&getname, "myapp", "CLIENT GETNAME");
 
-    // SELECT db
-    let sel = send_cmd(&mut s, &["SELECT", "0"]).await;
-    assert_contains(&sel, "OK", "SELECT 0");
+    // SELECT db (requires key on DB 0)
+    let sel = send_cmd(&mut s, &["SELECT", "0", "KEY", "test-admin"]).await;
+    assert_contains(&sel, "OK", "SELECT 0 with key");
 
     // QUIT should close connection after sending OK
     let quit = send_cmd(&mut s, &["QUIT"]).await;
@@ -279,7 +290,11 @@ async fn test_02_strings_and_expiry() {
 
     let ex0 = send_cmd(&mut s, &["EXISTS", "user:1"]).await;
     assert_contains(&ex0, "0", "EXISTS after DEL");
-
+    
+    // DEL non-existent should return 0
+    let del0 = send_cmd(&mut s, &["DEL", "user:1"]).await;
+    assert_contains(&del0, "0", "DEL user:1 when not exists -> 0");
+    
     // INCR behavior
     let i1 = send_cmd(&mut s, &["INCR", "count"]).await;
     assert_contains(&i1, "1", "INCR new key -> 1");
@@ -501,11 +516,11 @@ async fn test_07_age_stateless_suite() {
     let mut s = connect(port).await;
 
     // GENENC -> [recipient, identity]
-    let gen = send_cmd(&mut s, &["AGE", "GENENC"]).await;
+    let genenc = send_cmd(&mut s, &["AGE", "GENENC"]).await;
     assert!(
-        gen.starts_with("*2\r\n$"),
+        genenc.starts_with("*2\r\n$"),
         "AGE GENENC should return array [recipient, identity], got:\n{}",
-        gen
+        genenc
     );
 
     // Parse simple RESP array of two bulk strings to extract keys
@@ -520,7 +535,7 @@ async fn test_07_age_stateless_suite() {
         let ident = lines.next().unwrap_or("").to_string();
         (recip, ident)
     }
-    let (recipient, identity) = parse_two_bulk_array(&gen);
+    let (recipient, identity) = parse_two_bulk_array(&genenc);
     assert!(
         recipient.starts_with("age1") && identity.starts_with("AGE-SECRET-KEY-1"),
         "Unexpected AGE key formats.\nrecipient: {}\nidentity: {}",
@@ -591,7 +606,7 @@ async fn test_08_age_persistent_named_suite() {
 
     // AGE LIST
     let lst = send_cmd(&mut s, &["AGE", "LIST"]).await;
-    assert_contains(&lst, "encpub", "AGE LIST label encpub");
+    // After flattening, LIST returns a flat array of managed key names
     assert_contains(&lst, "app1", "AGE LIST includes app1");
 }