Use case
An MCP memory server for AI coding tools.
Cosmos exposes local memory and code intelligence through MCP tools that Claude Code, Cursor, Cline, Windsurf, Gemini CLI, and future clients can call.
- Protocol
- MCP
- Tools
- Brain, code, project lessons, context bundles
- Goal
- Less repeated context, fewer broad scans
What the MCP server provides
- Project structure and code symbols.
- Relevant files and past fixes for natural-language tasks.
- Memory search across notes, decisions, and saved lessons.
Why it is useful
MCP gives AI clients a structured way to ask for the right local context. Cosmos focuses on making those answers small, scoped, and auditable.