What vector databases are supported?

Claude Context supports Milvus (self-hosted) and Zilliz Cloud (managed) as vector database backends.

What embedding providers can I use?

Claude Context supports OpenAI, VoyageAI, Ollama, and Gemini as embedding providers.

Which AI coding agents are compatible?

Claude Context works with Claude Code, Cursor, Windsurf, VS Code, Cline/Roo Code, and any MCP-compatible client.

How does incremental indexing work?

Claude Context uses Merkle trees to detect which files have changed and only re-indexes those files, keeping the index up-to-date efficiently.

claude-context

Name: claude-context
Author: Zilliz

by Zilliz

communitystdiogithub

View all implementations

6.1kstars|538forks

View on GitHub Docs

Works withClaude DesktopCursorVS CodeWindsurf

About

Claude Context is an open-source MCP server by Zilliz that enables AI coding agents to perform semantic code searches across entire codebases using hybrid search (BM25 + dense vectors), achieving ~40% token reduction while providing deep, relevant context.

Features

Hybrid search combining BM25 and dense vector retrieval for natural language code queries
~40% token reduction by retrieving only relevant code snippets
Incremental indexing using Merkle trees for efficient re-indexing of changed files
AST-based code chunking for context-aware code splitting
Support for 14+ programming languages including TypeScript, Python, Java, Go, Rust
Multiple embedding provider support: OpenAI, VoyageAI, Ollama, Gemini
Milvus/Zilliz Cloud vector database backend
VSCode extension available as Semantic Code Search

Use Cases

AI-assisted developers

Enable coding agents to search entire codebases semantically, reducing token costs while maintaining deep context

Large codebase teams

Index massive repositories and let AI agents find relevant code sections without loading entire files

DevOps engineers

Quickly search across infrastructure code repositories using natural language queries

What This Server Can Do

MCP servers expose three types of capabilities to AI clients. Here's what claude-context supports.

Tools

Supported4 tools

Functions your AI client can call to perform actions — like querying a database, creating a file, or calling an API.

How to use: Tools run automatically when your AI client decides they're needed. Ask your AI assistant to perform a task, and it will invoke the right tool.

Resources

Not available

This server does not expose read-only data resources.

Prompts

Not available

This server does not provide pre-built prompt templates.

For the full list of available tools, resources, and prompts, check the README on GitHub.

Available Tools (4)

index_codebase

Indexes a directory for hybrid search. Creates or updates the search index for the specified project.

search_code

Executes natural language queries against the indexed codebase to find relevant code snippets.

clear_index

Deletes the search index for a specific project, removing all indexed data.

get_indexing_status

Provides progress percentage and completion status of the current indexing operation.

Alternative Implementations

Other MCP servers for the same service.

View all on service page

MCP Toolbox for Databases

official75/100

15.1k