Apppublic

Chris4K/agent-memory

sourceHugging Faceupdated 6mo agoView on Hugging Face

1likes

App README

🧠 AGENT MEMORY — Multi-tier Memory MCP

Memory tiers: episodic · semantic · procedural · working (TTL)

MCP — Claude Desktop

json

{
  "mcpServers": {
    "agent-memory": {
      "command": "npx",
      "args": ["-y", "mcp-remote", "https://chris4k-agent-memory.hf.space/mcp/sse"]
    }
  }
}

MCP Tools

Tool	Description
`memory_store`	Store a memory with tier, tags, importance
`memory_search`	Full-text search across all memories
`memory_recall`	Get memory by ID
`memory_update`	Update content, tags, importance
`memory_forget`	Delete a memory
`memory_list`	List by tier/tag/agent
`memory_stats`	Counts per tier

REST API

GET    /api/memories?tier=semantic&tag=jarvis
GET    /api/memories/search?q=GDPR&tier=all
GET    /api/memories/{id}
POST   /api/memories     {"content":"...","tier":"episodic","tags":[],"importance":7}
PATCH  /api/memories/{id}
DELETE /api/memories/{id}
GET    /api/stats

Agent Usage

python

import requests
BASE = "https://chris4k-agent-memory.hf.space"

# Store a memory
requests.post(f"{BASE}/api/memories", json={
    "content": "User prefers async Python with FastAPI for all backend services.",
    "tier": "semantic", "tags": ["preference","python"], "agent": "jarvis", "importance": 8
})

# Search
r = requests.get(f"{BASE}/api/memories/search?q=FastAPI&limit=5")
print(r.json())

Chris4K · ki-fusion-labs.de