Semantic Search MCP

MCP server that adds semantic search over markdown files to Claude Code. Find related notes by meaning, not just keywords. Detect duplicates before creating new notes.

Installation

From GitHub:

claude mcp remove -s project semantic-search
claude mcp add -s project semantic-search \
--env CONTENT_PATH=/path/to/your/content \
-- \
uvx --from git+https://github.com/bborbe/semantic-search-mcp semantic-search-mcp serve

Local development:

claude mcp remove -s project semantic-search
claude mcp add -s project semantic-search \
--env CONTENT_PATH=/path/to/your/content \
-- \
uvx --reinstall --from /path/to/semantic-search-mcp semantic-search-mcp serve

Replace /path/to/your/content with your markdown directory (e.g., Obsidian vault path).

Multiple Directories

Index multiple directories by separating paths with commas:

--env CONTENT_PATH=/path/to/vault1,/path/to/vault2,/path/to/docs

All directories are indexed together and searched as one unified index.

Usage

Ask Claude:

• "Do I have notes about X?" → searches related notes
• "Find similar notes to Y" → finds semantically related content
• Before creating new notes → Claude checks for duplicates automatically

Tools

• search_related(query, top_k=5) - Find semantically related notes
• check_duplicates(file_path) - Detect duplicate/similar notes

How It Works

First run downloads a small embedding model (~90MB) and indexes your markdown files (<1s for typical vaults). Each Claude Code session gets its own index in /tmp/ that auto-updates when files change. Multiple sessions work independently without conflicts.

License

This project is licensed under the BSD 2-Clause License - see the LICENSE file for details.