Claude Code forgets everything between sessions. CogmemAi fixes that with an MCP server that gives it real persistent memory. Semantic search finds relevant memories by meaning. Ai-powered extraction automatically identifies facts worth remembering. Compaction recovery hooks reload your context seamlessly when Claude Code compresses your conversation. One command setup: npx cogmemai-mcp setup Works with Claude Code, Cursor, Windsurf, Cline, and Continue. Free tier: 1,000 memories.
I built this because I was spending the first 5 minutes of every Claude Code
session re-explaining my project architecture. CLAUDE.md helps but tops out
at ~150 lines and gets unreliable after context compaction.
CogmemAi stores facts about your codebase server-side and retrieves them via
semantic search. The compaction recovery was the hardest part — Claude Code's
transcript format is undocumented, so getting the hooks right took real
iteration.
After weeks of daily use, the difference is night and day. Happy to answer
questions.
Report
No reviews yetBe the first to leave a review for CogmemAi