Official Model Context Protocol (MCP) Server for Thordata.
Give LLMs (Claude, Cursor) the ability to search the web, scrape data, and perform complex tasks.
This server exposes Thordata's AI infrastructure as callable tools for AI models:
- 🔍 Search: Real-time Google search (
google_search). - 📖 Read: Turn any webpage (including JS-heavy SPAs) into clean Markdown (
read_url). - 🗺️ Maps: Extract business details from Google Maps links (
get_google_maps_details). - 🎥 Media: Retrieve YouTube/Instagram metadata (
get_youtube_video_info).
-
Clone this repository:
git clone https://github.com/Thordata/thordata-mcp-server.git cd thordata-mcp-server -
Install dependencies:
# Recommend using uv for speed, or pip pip install .
-
Add to your
claude_desktop_config.json:- MacOS:
~/Library/Application Support/Claude/claude_desktop_config.json - Windows:
%APPDATA%\Claude\claude_desktop_config.json
{ "mcpServers": { "thordata": { "command": "python", "args": ["-m", "thordata_mcp.main"], "cwd": "/absolute/path/to/thordata-mcp-server", "env": { "THORDATA_SCRAPER_TOKEN": "your_token", "THORDATA_PUBLIC_TOKEN": "your_public_token", "THORDATA_PUBLIC_KEY": "your_public_key" } } } } - MacOS:
-
Set up environment:
cp .env.example .env # Edit .env with your tokens -
Run the MCP Inspector:
mcp dev src/thordata_mcp/main.py
docker build -t thordata-mcp .
docker run -i --env-file .env thordata-mcpIf you use Smithery, you can install this server directly:
npx -y @smithery/cli install @thordata/thordata-mcp-server --client claudeMIT License.