# Mycelium Agents – Architecture Overview

## Personal Agents
- Every individual has a **Personal Agent (PA)**.  
- The PA manages tasks, workflows, and digital life.  
- It acts as the secure, sovereign interface to data and AI services.  

## Worker Agents
- Specialized agents that execute tasks such as research, coding, analytics, or operations.  
- Activated and coordinated by Personal Agents.  

## LLMs (Brains)
- Large Language Models provide reasoning, language understanding, and advanced knowledge.  
- Accessed by Personal Agents and Worker Agents when higher cognitive capacity is needed.  

## Network of Intelligence
- Agents communicate securely with each other.  
- Private, encrypted peer-to-peer connections.  
- Enables collaboration within and across organizations.  

## Data Control
- All data stays under user or organization control.  
- Agents use sovereign storage and secure communication.  



## Architecture Diagram

```mermaid
%%{init: {'theme':'dark'}}%%
graph TD

    H[Human User] --> PA[Personal Agent]

    PA --> LLM[(LLM / Brain)]
    PA --> WA1[Worker Agent: Research]
    PA --> WA2[Worker Agent: Coding]
    PA --> WA3[Worker Agent: Analytics]

    %% Agents can collaborate
    WA1 --- WA2
    WA2 --- WA3
    PA --- PAX[Other Personal Agent]

```