# m3 Memory

> Local-first, MCP-native memory layer for coding agents (Claude Code, Codex, Gemini CLI, Antigravity, OpenCode, OpenClaw) with hybrid search and chat capture.

- id: `m3-memory`
- category: Tools (`tools`)
- url: https://github.com/skynetcmd/m3-memory
- repo: https://github.com/skynetcmd/m3-memory
- docs: https://github.com/skynetcmd/m3-memory/blob/main/docs/GETTING_STARTED.md
- status: active
- pricing: open-source
- license: Apache-2.0
- tags: memory, mcp, local-first, coding-agents, python
- related: MCP Memory Service (https://indexagentica.com/entries/mcp-memory-service/); Basic Memory (https://indexagentica.com/entries/basic-memory/); agentmemory (https://indexagentica.com/entries/agentmemory/); OmniMemEval (https://indexagentica.com/entries/omnimemeval/)
- sources: https://github.com/skynetcmd/m3-memory , https://github.com/skynetcmd/m3-memory/releases , https://pypi.org/project/m3-memory/
- added: 2026-10-27
- updated: 2026-10-27
- last verified: 2026-10-27
- featured in: Memory servers for coding agents compared: m3 Memory vs agentmemory vs Basic Memory vs MCP Memory Service (https://indexagentica.com/compare/coding-agent-memory-servers/)
- maintainer: skynetcmd
- submitted by: agentica-curator
- page: https://indexagentica.com/entries/m3-memory/
- json: https://indexagentica.com/api/entries/m3-memory.json

`pipx install m3-memory`, then `m3 setup` detects installed agents and wires m3 into each (MCP config plus session-start and chat-capture hooks), and `m3 doctor` checks the install. Stores facts, decisions, preferences and conversation history locally with hybrid search (SQLite FTS5 + vectors + MMR); works fully offline, needs no cloud account or API key, and can optionally use a local LLM (Ollama, LM Studio) to summarize and link memories. The project reports 99.2% LongMemEval-S retrieval at k=8 (self-reported). Python 3.12+; PyPI `m3-memory` 2026.10.25.0.