Updated April 2026

Install Semantic Search MCP Server

Pick your client, copy the command, done.

Local setup required. This server has to be cloned and prepared on your machine before you register it in Claude Code.

Run this once to clone and prepare the server before adding it to Claude Code.

Run in terminal

pip install -e .

After the local setup is done, run this command to point Claude Code at the built server.

Run in terminal

claude mcp add semantic-search -- node "<FULL_PATH_TO_SEMANTIC_SEARCH_MCP>/dist/index.js"

Replace <FULL_PATH_TO_SEMANTIC_SEARCH_MCP>/dist/index.js with the actual folder you prepared in step 1.

Environment Variables

Set these before running Semantic Search MCP Server.

VariableDescriptionRequired

REPO_PATHPath to the repository to searchNo

SEARCHER_TYPESearcher implementation to use (default: sgr_gemini_flash_lite)No

CLAUDE_API_KEYAPI key for Claude-based searchersNo

GOOGLE_API_KEYAPI key for Gemini-based searchersNo

OPENAI_API_KEYAPI key for OpenAI-based searchersNo

Once configured, Semantic Search MCP Server gives your AI agent access to:

semantic_searchPerforms a semantic search over the codebase using natural language queries.

query

After setup, try these prompts with your AI agent:

→Find code snippets related to user authentication logic in the current repository.

→Search for where the database connection is initialized and configured.

→Locate the implementation of the file upload handler.

→Find all instances where the API rate limiting logic is applied.

Prerequisites & system requirements

An MCP-compatible client (Claude Code, Cursor, Windsurf, Claude Desktop, or Codex)
Python 3.8+ with pip installed

Save the docs, env vars, and workflow around Semantic Search MCP Server in Conare so Claude Code, Codex, and Cursor remember it next time.