Framework Integrations

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

Table of Contents

1. Using with AMFS SaaS
2. CrewAI
3. LangGraph
4. LangChain
5. AutoGen
6. Building Your Own Integration

---

Using with AMFS SaaS

When using AMFS as a hosted service (SaaS), all integrations should connect via the HTTP API with an API key. Set these two environment variables before running your agents:

export AMFS_HTTP_URL="https://amfs-login.sense-lab.ai"
export AMFS_API_KEY="amfs_sk_your_key_here"

With these set, AgentMemory auto-detects the HTTP adapter and routes all operations through the authenticated API. This applies to CrewAI, LangGraph, LangChain, AutoGen, and any custom integration.

Never use AMFS_POSTGRES_DSN for external agents in multi-tenant mode. Always use AMFS_HTTP_URL + AMFS_API_KEY to ensure tenant isolation.

See the SaaS Connection Guide for detailed setup instructions.

---

CrewAI

Add AMFS tools to your CrewAI agents:

pip install amfs-crewai

from amfs import AgentMemory
from amfs_crewai import AMFSTool

mem = AgentMemory(agent_id="crewai-agent")
tools = AMFSTool(mem).tools()
# Returns [AMFSReadTool, AMFSWriteTool, AMFSListTool]

# Pass to your CrewAI agent
from crewai import Agent

agent = Agent(
    role="Researcher",
    goal="Analyze the codebase",
    tools=tools,
)

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:

pip install amfs-langgraph

from amfs import AgentMemory
from amfs_langgraph import AMFSCheckpointer

mem = AgentMemory(agent_id="graph-agent")
checkpointer = AMFSCheckpointer(mem)

# Pass to your LangGraph graph builder

The checkpointer persists graph state to AMFS, enabling cross-session graph execution.

---

LangChain

Use AMFS as chat memory for LangChain:

pip install amfs-langchain

from amfs import AgentMemory
from amfs_langchain import AMFSChatMemory

mem = AgentMemory(agent_id="chat-agent")
chat_memory = AMFSChatMemory(mem, session_key="conv-123")

# Save conversation context
chat_memory.save_context(
    {"input": "What's the retry policy?"},
    {"output": "We use exponential backoff with max 3 retries."},
)

# Load conversation history
history = chat_memory.load_memory_variables({})

---

AutoGen

Use AMFS as a memory store for AutoGen agents:

pip install amfs-autogen

from amfs import AgentMemory
from amfs_autogen import AMFSMemoryStore

mem = AgentMemory(agent_id="autogen-agent")
store = AMFSMemoryStore(mem)

# Store and retrieve data
store.add("user_prefs", {"theme": "dark"})
prefs = store.get("user_prefs")

---

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)

class MyFrameworkMemory:
    def __init__(self, mem: AgentMemory, scope: str = "default"):
        self._mem = mem
        self._scope = scope

    def save(self, key: str, data: dict):
        self._mem.write(self._scope, key, data)

    def load(self, key: str):
        entry = self._mem.read(self._scope, key)
        return entry.value if entry else None