LangGraph + Akasha Pipeline¶

A complete example of 4 AI agents sharing knowledge through Akasha — zero orchestration.

Architecture¶

graph LR
    R[Researcher Agent] -->|write discoveries| M[Akasha Memory]
    M -->|read data| A[Analyst Agent]
    A -->|write analysis| M
    M -->|read report| W[Writer Agent]
    W -->|write draft| M
    M -->|read draft| C[Critic Agent]
    C -->|write feedback| M

    style M fill:#6c5ce7,stroke:#a29bfe,color:#fff

How It Works¶

1. Researcher gathers data → writes to memory/working/research/*
2. Analyst reads research → writes analysis to memory/episodic/analysis/*
3. Writer reads analysis → writes draft to memory/working/draft
4. Critic reads draft → writes feedback to memory/semantic/feedback/*

No agent talks to another directly. They coordinate through shared memory.

Run It¶

cd examples/langgraph-memory
pip install -r requirements.txt
python pipeline.py

Source Code¶

See the full example on GitHub: examples/langgraph-memory/