memory-mesh

[](https://pypi.org/project/memory-mesh/) [](#) [](#) [](#)

> Give Claude, ChatGPT and Gemini shared memory.

memory-mesh is a local-first MCP server and REST API that lets every AI assistant on your machine read and write to the same memory store — with automatic conflict detection when two agents disagree.

---

No shared keys. No coordination. Just two agents writing what they know — and memory-mesh catching when they contradict each other.

# Claude saves a fact
curl -s -X POST http://localhost:8765/v1/memories \
  -H "Content-Type: application/json" \
  -d '{"content": "The Eiffel Tower is 300 metres tall", "agent_id": "claude"}'

# Gemini saves a contradicting fact
curl -s -X POST http://localhost:8765/v1/memories \
  -H "Content-Type: application/json" \
  -d '{"content": "The Eiffel Tower is 330 metres tall", "agent_id": "gemini"}'

The second response includes:

{
  "memory": { "...": "..." },
  "conflicts": [{
    "id": "abc-123",
    "trigger": "similarity",
    "memory_a_id": "...",
    "memory_b_id": "..."
  }]
}

Resolve it interactively:

$ memory-mesh conflicts

============================================================
Conflict ID: abc-123  [similarity]

  A [...] (claude):
    The Eiffel Tower is 300 metres tall

  B [...] (gemini):
    The Eiffel Tower is 330 metres tall

  Which is correct? [a/b/skip]: a
  Resolved: claude's memory wins.

---

pip install memory-mesh

---

Add to your claude_desktop_config.json:

{
  "mcpServers": {
    "memory-mesh": {
      "command": "memory-mesh",
      "args": ["mcp"]
    }
  }
}

Claude can now call save_context, `search