Local-first agent memory MCP — Claude Code, Cursor, OpenAI Codex, and Gemini CLI share one SQLite brain. 6-layer structured memory (goal/context/emotion/implementation/caveat/learning), drift detection with pre-action re-injection guard, and an AST-aware diff cache saving 50-99% tokens on file re-reads. npx linksee-memory-setup — one command.
Overview
Linksee Memory
Your agent forgets everything when a session ends. Worse — it silently drifts from what you decided last week.
Linksee Memory is a local-first cross-LLM memory MCP: one SQLite file that Claude Code, Cursor, Windsurf, OpenAI Codex, and Gemini CLI all read from.
Features
- 6-layer structured memory — goal / context / emotion / implementation / caveat / learning. Pain records (caveats) are never auto-forgotten
- Drift detection (Intent Datadog) — declare decisions as anchors, get flagged when reality diverges without a recorded resolution
- Re-injection guard — re-surfaces locked decisions BEFORE the agent acts (Claude Code hooks)
- read_smart — AST-aware diff cache, 86% tokens saved on a typical re-read, 99% on unchanged files
- Full MCP surface — Tools + Resources + Prompts + Sampling + Roots + Elicitation
Install
npx linksee-memory-setup
Docs & onboarding: https://linksee-site.vercel.app
Server Config
{
"mcpServers": {
"linksee": {
"command": "npx",
"args": [
"-y",
"linksee-memory"
]
}
}
}