Key takeaways
- CLI-first semantic code search powered by ChromaDB and AST parsing, with MCP integration for AI coding assistants
- Analysis capabilities beyond search: complexity analysis, dead code detection, SARIF output for CI integration
- Interactive chat mode with iterative refinement (up to 30 queries) and advanced reasoning mode for architectural analysis
- 24 stars, Python. Early-stage but feature-rich for its size — combines vector search with static analysis
FAQ
What is mcp-vector-search?
A CLI-first semantic code search tool powered by ChromaDB vector storage and AST parsing. Provides MCP integration, complexity analysis, dead code detection, and an interactive chat mode for codebase exploration.
How does it compare to other code search MCP tools?
More analysis-oriented than Octocode (which focuses on cross-repo search). Smaller but combines vector search with static analysis features like complexity scoring and dead code detection.
Overview
mcp-vector-search is a CLI-first semantic code search tool powered by ChromaDB and AST parsing. It combines vector-based semantic search with static analysis capabilities (complexity analysis, dead code detection) and exposes everything via MCP for AI coding assistants.
The interactive chat mode supports iterative refinement with up to 30 queries and an advanced reasoning mode for architectural analysis. Outputs in JSON, SARIF (for CI), and markdown formats.
Key stats: 24 stars, Python. Created August 2025.
Competitive Position
Strengths: Combines vector search with static analysis. SARIF output for CI. Interactive chat mode. Feature-rich for its size.
Weaknesses: Very small community (24 stars). Early stage. ChromaDB dependency adds setup complexity.
Research by Ry Walker Research