This is an early release preview. You may encounter bugs.
Claude Context logo
Unclaimed

Tool coding memory

Claude Context

MCP plugin indexing a codebase into a vector database for semantic search from AI coding agents

B 65/100 GitHub score ? This grade is derived from GitHub signals, not user votes. Open for the full breakdown.
No votes yet

01 / About

What Claude Context is.

Claude Context is an MCP plugin that indexes a codebase into a vector database and exposes semantic code search to AI coding agents such as Claude Code. Instead of loading whole directories into the model's context for each request, the agent runs natural-language queries against the index and receives only the related code chunks. A controlled evaluation reports about 40% fewer tokens at equivalent retrieval quality.

Indexing chunks source files using an abstract syntax tree (AST) splitter, with a character-based splitter as fallback, embeds the chunks, and stores them in Milvus or Zilliz Cloud. Search combines BM25 keyword matching with dense vector retrieval. A Merkle tree tracks file changes so re-indexing covers only modified files. The MCP server uses stdio transport and works with any MCP-compatible client, with configuration examples for Claude Code, Codex CLI, Gemini CLI, Qwen Code, Cursor, Windsurf, VS Code, Cline, Roo Code, Augment, Zencoder, Cherry Studio, and Void.

The project is a TypeScript monorepo (Node.js 20+) with three packages: @zilliz/claude-context-core, the indexing and search engine usable directly in applications; @zilliz/claude-context-mcp, the MCP server; and a VS Code extension, Semantic Code Search.

Features

  • MCP tools: indexcodebase, searchcode, clearindex, and getindexing_status
  • Hybrid search: BM25 plus dense vector retrieval over the indexed chunks
  • AST-based chunking: syntax-aware code splitting with automatic fallback to a character-based splitter
  • Incremental indexing: Merkle trees identify changed files so only those are re-indexed
  • Embedding providers: OpenAI, VoyageAI, Ollama, and Gemini, with configurable models such as text-embedding-3-large or voyage-code-3
  • Vector databases: Milvus or the managed Zilliz Cloud service
  • Language coverage: TypeScript, JavaScript, Python, Java, C++, C#, Go, Rust, PHP, Ruby, Swift, Kotlin, Scala, and Markdown
  • Customizable rules: file extensions, ignore patterns, and embedding models are configurable
  • Core package API: Context, MilvusVectorDatabase, and embedding classes for building your own indexing and search flows
  • VS Code extension: semantic search and navigation inside the editor

02 / Discussion CREDIBILITY-GATED

Discussion

Reading is open to everyone. Posting and voting need a verified identity or a GitHub grade of B or higher.

  • No discussions yet.

04 / Build

Build with Claude Context.

Browse the catalogue for frameworks, tools, and harnesses, each scored on real GitHub credibility.

Get Claude Context →

Browse the catalogue