http-generic

HTTP Generic Integration Guide

Connect any agent framework or HTTP client to mcp-memory-service using the REST API.

Server Setup

pip install mcp-memory-service
MCP_ALLOW_ANONYMOUS_ACCESS=true memory server --http
# Running at http://localhost:8000

With API key authentication:

MCP_API_KEY=your-secret-key memory server --http
# Include header: Authorization: Bearer your-secret-key

All REST Endpoints

Method	Endpoint	Description
POST	/api/memories	Store a new memory
GET	/api/memories	List memories (paginated)
GET	/api/memories/{hash}	Get memory by hash
PATCH	/api/memories/{hash}	Update memory tags/type/metadata
DELETE	/api/memories/{hash}	Delete a memory
POST	/api/memories/search	Semantic + keyword search
GET	/api/memories/tags	List all tags with counts
GET	/api/health	Health check
GET	/api/stats	Storage statistics
POST	/api/consolidation/run	Trigger memory consolidation
GET	/api/consolidation/status	Consolidation status
GET	/api/graph/associations/{hash}	Get memory associations
POST	/api/graph/associations	Store association between memories
GET	/api/analytics/relationship-types	Graph relationship type stats
GET	/sse/events	Server-Sent Events stream

Authentication Patterns

# Anonymous (MCP_ALLOW_ANONYMOUS_ACCESS=true)
curl http://localhost:8000/api/memories

# API key
curl -H "Authorization: Bearer $MCP_API_KEY" http://localhost:8000/api/memories

# OAuth (see docs/oauth/)
curl -H "Authorization: Bearer $ACCESS_TOKEN" http://localhost:8000/api/memories

Python (httpx) Examples

Store a memory

import httpx

BASE_URL = "http://localhost:8000"

async def store_memory(content: str, tags: list[str], agent_id: str | None = None) -> dict:
    headers = {"Content-Type": "application/json"}
    if agent_id:
        headers["X-Agent-ID"] = agent_id

    async with httpx.AsyncClient() as client:
        response = await client.post(
            f"{BASE_URL}/api/memories",
            json={"content": content, "tags": tags},
            headers=headers,
        )
        response.raise_for_status()
        return response.json()

# Usage
result = await store_memory(
    content="API rate limit is 100 req/min for the Acme service",
    tags=["api", "rate-limit", "acme"],
    agent_id="researcher",
)
print(result["memory"]["content_hash"])