io.github.DiaaAj/a-mem-mcp

io.github.DiaaAj/a-mem-mcp·v0.2.1·Memory & RAG

Quality Score

/100

Self-evolving memory system for AI agents

Repository →

§01 Install

Claude Desktop (uvx)

{
  "mcpServers": {
    "a-mem-mcp": {
      "command": "uvx",
      "args": [
        "a-mem"
      ],
      "env": {
        "LLM_BACKEND": "openai",
        "LLM_MODEL": "gpt-4o-mini",
        "OPENAI_API_KEY": "<your-openai_api_key>",
        "EMBEDDING_MODEL": "all-MiniLM-L6-v2",
        "CHROMA_DB_PATH": "./chroma_db"
      }
    }
  }
}

§02 Environment variables

LLM_BACKEND

LLM backend to use (openai, ollama, sglang, or openrouter)

LLM_MODEL

LLM model name (e.g., gpt-4o-mini, llama2, etc.)

OPENAI_API_KEY

requiredsecret

OpenAI API key (required if LLM_BACKEND=openai)

EMBEDDING_MODEL

Sentence transformer model for embeddings

CHROMA_DB_PATH

ChromaDB storage directory path