producthunt

This MCP server exposing Product Hunt data plugs into any client that supports the Model Context Protocol (MCP). It fetches posts, collections, topics, users, votes, and comments from Product Hunt and presents them through MCP-compatible tool calls, enabling an agent or chat UI to query and browse data without direct API handling. You can use tools like get_post_details, get_posts, get_post_comments, get_collection, get_collections, get_topic, search_topics, get_user, get_viewer, and check_server_status to retrieve structured data and metadata. The server is designed to be fast and compatible with MCP clients such as Claude Desktop and Cursor, or any custom MCP client you build.

To use it, install the server, start it with your Product Hunt API token set as an environment variable, and configure your MCP client to point to the running server. A typical configuration maps a server name (like product-hunt) to the command that starts the MCP server and supplies the necessary token via environment variables. When integrated with Claude Desktop or Cursor, you’ll add an entry under mcpServers with the command and env settings so the client can authenticate and make requests through the MCP interface.

Prerequisites:

• Python 3.10 or newer
• pip (comes with Python)
• A Product Hunt API token (Developer Token) for authentication

Install from PyPI (recommended):

pip install product-hunt-mcp

Run locally (example):

export PRODUCT_HUNT_TOKEN=your_token_here
python -m product_hunt_mcp

Alternative Docker flow:

# Build the image (if a Dockerfile is provided)
docker build -t product-hunt-mcp .
# Run with token (interactive)
docker run -i --rm -e PRODUCT_HUNT_TOKEN=your_token_here product-hunt-mcp

If you prefer installing from source or using uv (optional for development), you can follow the repository’s guidance to install via uv and then run the module, but the standard usage is to rely on the Python package entry point as shown above.

Tips and notes:

• The PRODUCT_HUNT_TOKEN must be set in the environment of the MCP client or the server itself, depending on your deployment method. Do not hard-code tokens into configuration files.
• If Claude Desktop or Cursor cannot locate the python executable or the server binary, provide the full path to the executable in your client configuration.
• If you use Docker, ensure the container has access to network resources and that the host's token is provided via environment variables.
• The MCP tools expose many query parameters (e.g., get_posts supports topic, order, count, etc.). Refer to the tool definitions to craft precise requests.
• Rate limiting may occur from Product Hunt; your client can check status via check_server_status or get_api_rate_limits tools.