# RemembrallMCP

RemembrallMCP gives AI coding agents whole-codebase knowledge: a field-aware dependency graph of functions, classes, methods and fields, plus persistent memory that survives between sessions, exposed to agents over MCP.

RemembrallMCP builds a live dependency graph of a codebase with tree-sitter, tracking functions, classes, methods, fields and the call, import, defines, inherits and field-reference relationships between them across nine languages. An agent can ask what a change would break down to a single struct field and get an answer before it touches anything.

The project also keeps persistent memory across sessions: decisions, patterns and organizational knowledge that an agent can recall instead of rediscovering. Memory can be searched with hybrid semantic and full-text search, and bulk-imported from merged pull request descriptions or documentation.

It is written in Rust, stores its graph and memory in Postgres with pgvector, and exposes both the code graph and the memory store as tools over the Model Context Protocol.

Website: https://roboticforce.io/remembrall