Skip to content
Hot eventDeveloping

Project-M releases local MCP memory server for AI coding agents

1 reports1 reporting sourcesUpdated 15 min ago

Start here

AI digest

Project-M released an open-source, local-first MCP server that shares project state across Claude Desktop, OpenAI Codex and Google Antigravity, so context carries over when a developer switches between assistants. It compiles a four-tier memory hierarchy and serves a Tier 1 briefing card of under 450 tokens in under 2ms, backed by SQLite WAL storage and FTS5 BM25 search. The project says it runs entirely on the user's machine with no cloud dependencies and no paid embedding syncs or third-party vector databases, and is MIT-licensed. It ships 8 FastMCP tools, a CLI/REPL and a one-click installer for the three clients, with 102 tests passing. These performance and cost claims come from the project itself and have not been independently verified.

Generated from reports · updated 10 min ago

Report timeline

Follow the reports to see the event from different sides.

November 18
  1. Hacker News · AI
    Project-M launches a local shared-memory MCP server for AI coding agents

    Project-M is an open-source, local-first MCP server that shares project state across Claude Desktop, OpenAI Codex and Google Antigravity so context survives switching between assistants. It compiles a four-tier memory hierarchy, serving a Tier 1 briefing card under 450 tokens in under 2ms, backed by SQLite WAL storage and FTS5 BM25 search with no cloud dependencies. The repo ships 8 FastMCP tools, a CLI/REPL and a one-click installer for the three clients, with 102 tests passing.

Heat for this event

Not enough continuous observations yet to draw a trend.