In Memoria Memory Server

Example Use Cases

• 1

  Persistent, local memory of a codebase's patterns and architecture to produce context-aware suggestions across sessions

• 2

  Semantic search and file retrieval (AST-aware) to locate relevant code by meaning rather than keywords

• 3

  Pattern-consistent recommendations (naming, signatures, project organization) when generating or refactoring code

Description

In Memoria analyzes codebases (AST parsing + pattern learning) and exposes learned intelligence via the Model Context Protocol so AI assistants can access persistent memory across sessions. It extracts naming conventions, architectural preferences, and semantic relationships, stores structured and vector data locally (SQLite + SurrealDB), and offers 17 MCP tools for search, analysis, and recommendations. The server runs locally (Rust core + TypeScript MCP layer), integrates with tools like Copilot and Claude, and helps agents give context-aware, style-consistent suggestions without re-analyzing the repo every session.