Skip to content
  1. Home
  2. MCP servers
  3. Claude Context

Claude Context

Semantic code search over a whole codebase, backed by a vector database

What it is

An MCP server from Zilliz that indexes a repository into a Milvus/Zilliz vector store and lets the agent search it semantically, returning relevant chunks instead of whole files. Meant for large codebases where loading directories into context is too costly.

Good for

  • Index a repository for semantic search
  • Find code by describing what it does
  • Cut token use on large codebases

Install or run

npx @zilliz/claude-context-mcp@latest

Command as given in the project README. Read the repository before you run anything on your machine.

What you need

Node.js; embedding provider key (e.g. OPENAI_API_KEY); Milvus or Zilliz Cloud token

Worth knowingCode chunks are sent to an embedding provider and stored in a vector database, which is external in the default setup.

Sources

The summary is ours, written from the README. Spotted an error? See how we work.

More for coding

See the topic →
  • REA: Reverse Engineer Anything

    morluto/rea

    MCP server and CLI for agent-driven reverse engineering of apps and binaries

    MCP serverGitHub stars: 456k
  • Context7

    upstash/context7

    Version-specific library docs and code examples fetched into the agent

    MCP serverGitHub stars: 65k
  • Chrome DevTools for agents

    ChromeDevTools/chrome-devtools-mcp

    Lets a coding agent control and inspect a live Chrome via DevTools

    MCP serverGitHub stars: 55k
  • GitNexus

    abhigyanpatwari/GitNexus

    Indexes a codebase into a graph of symbols, calls and execution flows

    MCP serverGitHub stars: 48k