#1414th of 15 in MCP servers for developers
claude-context
MCP server adding hybrid semantic code search to coding agents
- Stars
- 12.6k
- License
- MIT
- Last commit
- Jul 2026
- Language
- TypeScript
5,645 on 25 Mar 202612,616 on 14 Oct 2026
From GitHub's daily star counts.
Checked on GitHub.
Zilliz is hiring: 11 open roles
Overview
Claude Context indexes a codebase into Milvus or Zilliz Cloud and exposes index_codebase, search_code, clear_index and get_indexing_status tools over stdio MCP. Search combines BM25 with dense vectors, chunks code with an AST splitter, and re-indexes only changed files using Merkle trees. It works with Claude Code, Codex CLI, Gemini CLI, Cursor and other MCP clients.
Who it is for: Developers giving Claude Code or other agents search over large repos
Strengths
- Hybrid BM25 plus dense vector search over an indexed repo
- Incremental re-indexing of changed files via Merkle trees
- Embedding choice: OpenAI, VoyageAI, Ollama or Gemini
- Also ships a core npm package and a VS Code extension
Weaknesses
- Needs a Milvus or Zilliz Cloud instance; no embedded vector store mentioned
- Default setup sends code to OpenAI for embeddings unless Ollama is configured
- README claims ~40% token reduction from its own evaluation only
- No tagged releases listed on the repository
What it needs
- no GPU
- Needs Node.js >= 20, Milvus or Zilliz Cloud, OpenAI API key (default embedding provider)
- Models: OpenAI, VoyageAI, Ollama, Gemini
Also in MCP servers for developers
See all 15| Rank | Project | Score |
|---|---|---|
| 1 | gitnexusCode knowledge graph exposed to coding agents through MCP tools | 89 out of 100 |
| 2 | codegraphLocal code knowledge graph served to coding agents over MCP | 81 out of 100 |
| 3 | chrome-devtools-mcpMCP server that lets coding agents drive and debug a live Chrome | 70 out of 100 |
| #4 | codebase-memory-mcpLocal MCP server that indexes code into a queryable knowledge graph | 69 out of 100 |
| #5 | github-mcp-serverMCP server exposing GitHub repos, issues, PRs and Actions to AI tools | 68 out of 100 |