Code Memory is an MCP server designed to combat the forgetfulness of AI coding assistants. It offers a persistent knowledge base for your codebase, enabling AI to understand context between sessions, recall architectural decisions, leverage historical discussions, and comprehend code relationships. By integrating semantic search, Git history analysis, dependency graphs, session learning, and persistent knowledge storage, Code Memory ensures AI assistants never lose critical context, significantly enhancing their effectiveness and productivity in development workflows.
Key Features
01Semantic Search (local embeddings, offline)
02Git History Analysis (architectural decisions, rationale)
03Dependency Graphs (across 16 languages)
04Session Learning (patterns from coding sessions)
05Persistent Knowledge (remember important facts)
060 GitHub stars
Use Cases
01Retrieve code based on its meaning, not just keywords, to quickly find relevant sections.
02Understand the rationale behind past architectural and refactoring decisions from Git history.
03Analyze code relationships and dependencies to grasp system structure and impact of changes.
