---
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)**.