Skip to main content

Framework Integrations

AMFS integrates with popular multi-agent frameworks so you can add persistent memory to your existing workflows.

Setup

Set your connection env vars — all framework integrations pick them up automatically:

CrewAI

Add AMFS tools to your CrewAI agents:
The tools give your CrewAI agents the ability to read, write, and list memory entries as part of their task execution.

LangGraph

Use AMFS as a checkpointer for LangGraph:
The checkpointer persists graph state to AMFS, enabling cross-session graph execution.

LangChain

Use AMFS as chat memory for LangChain:

AutoGen

Use AMFS as a memory store for AutoGen agents:

Building Your Own Integration

All integrations are thin wrappers around the AgentMemory API. To build your own:
  1. Accept an AgentMemory instance in your constructor
  2. Map your framework’s memory/tool API to mem.read(), mem.write(), mem.list(), etc.
  3. Use entity_path to scope data (e.g., per-conversation, per-task, per-agent)