88 lines
2.9 KiB
Markdown
88 lines
2.9 KiB
Markdown
---
|
||
sidebar_position: 3
|
||
---
|
||
|
||
# Agent Use Cases
|
||
|
||
This page adapts the **Agent Use Cases** section from the Mycelium site into documentation form.
|
||
|
||
## What the Framework Enables
|
||
|
||
> *“The framework gives you full control over where agents live, how they connect, and what data they use.”*
|
||
|
||
Agents are intended to be:
|
||
|
||
- **Location‑aware** – you decide which nodes or VDCs they inhabit.
|
||
- **Network‑aware** – connectivity and topology are visible and controllable.
|
||
- **Data‑aware** – data stays where it should, with explicit movement.
|
||
|
||
## Example Use Cases
|
||
|
||
### Run Agents on Your Own Hardware
|
||
|
||
Deploy agents directly on:
|
||
|
||
- Laptops and workstations.
|
||
- Homelabs and edge boxes.
|
||
- VDCs and clusters you control.
|
||
|
||
You are not tied to any single cloud or vendor; agents execute where it makes the most sense for latency, privacy, or cost.
|
||
|
||
### Connect Agents Over the Mycelium Network
|
||
|
||
Use **Mycelium Network** as the secure fabric between agents.
|
||
|
||
- Private overlay addressing across homes, offices, and datacenters.
|
||
- Encrypted paths between nodes in different countries or environments.
|
||
- One consistent address space spanning local and remote infrastructure.
|
||
|
||
### Keep Data and Memory Private by Default
|
||
|
||
Agents are designed to:
|
||
|
||
- Use local datasets, tools, prompts, and embeddings.
|
||
- Avoid sending sensitive context to external AI APIs by default.
|
||
- Respect policies for which data may be shared and where.
|
||
|
||
This is especially important for regulated sectors and internal systems.
|
||
|
||
### Build Workflows Across Cloud + Edge
|
||
|
||
Orchestrate multi‑node workflows that span:
|
||
|
||
- Edge clusters near data sources.
|
||
- Central VDCs for heavier compute or aggregation.
|
||
- Personal devices for interaction and control.
|
||
|
||
Examples:
|
||
|
||
- Real‑time processing at the edge, with summarized results aggregated in a VDC.
|
||
- Agents coordinating tasks between office locations and remote workers.
|
||
- Data pipelines that never leave your chosen infrastructure.
|
||
|
||
### Operate in Regulated Contexts
|
||
|
||
Run agents where strict requirements apply:
|
||
|
||
- Healthcare, finance, public sector, and other regulated domains.
|
||
- Environments with **data residency** constraints.
|
||
- Workloads requiring verified identity and controlled connectivity.
|
||
|
||
The combination of **local execution**, **sovereign infrastructure**, and **encrypted mesh networking** is intended to make regulatory compliance more tractable.
|
||
|
||
### Blend Local and Remote Intelligence
|
||
|
||
Not all tasks need to run on the same node:
|
||
|
||
- Lightweight agents might run locally for responsiveness and interactivity.
|
||
- Heavier workloads can be scheduled to trusted remote nodes (e.g. GPU clusters).
|
||
- Data movement is explicit and policy‑driven.
|
||
|
||
This supports hybrid strategies where you balance privacy, performance, and cost.
|
||
|
||
---
|
||
|
||
## Looking Ahead
|
||
|
||
These use cases are **illustrative** and may expand as the Agent Framework matures. For how to prepare your infrastructure today, see **[Getting Ready for Agents](/ai-agent-framework/getting-ready)**.
|