tavily -sse
Basic remote SSE MCP server for Tavily Search.
claude mcp add --transport stdio noesamaille-tavily-mcp-sse python -m uvicorn server:app --host 0.0.0.0 --port 8000 --reload \ --env API_KEY="<YOUR_API_KEY>" \ --env TAVILY_API_KEY="<YOUR_TAVILY_API_KEY>"
How to use
This MCP server exposes a FastAPI-based interface that acts as a remote SSE MCP server for interacting with the Tavily API. It requires two API keys: API_KEY for accessing the server's endpoints and TAVILY_API_KEY for authenticating with the Tavily API itself. Once running, you can use the provided endpoints to perform Tavily searches and content extraction. The server is wrapped with the fastapi-mcp framework to enable remote SSE streaming capabilities, so you can integrate it into an MCP workflow and receive updates as events. To query, send a POST request to /tavily-search with a TavilySearchRequest payload to perform a search, or to /tavily-extract with a TavilyExtractRequest to extract content from URLs. Be sure to set the required environment variables before starting the server.
How to install
Prerequisites:
- Python 3.8+ installed on your system
- Access to install Python packages (pip)
Installation steps:
-
Create a virtual environment (optional but recommended): python -m venv venv source venv/bin/activate # On Windows use: venv\Scripts\activate
-
Install dependencies from requirements.txt: pip install -r requirements.txt
-
Set up environment variables (before running the server): export API_KEY=your_api_key # Linux/macOS set API_KEY=your_api_key # Windows export TAVILY_API_KEY=your_tavily_api_key set TAVILY_API_KEY=your_tavily_api_key
-
Run the server (the command used by MCP config): python -m uvicorn server:app --host 0.0.0.0 --port 8000 --reload
-
Verify the server is running by visiting http://127.0.0.1:8000
Additional notes
Tips and considerations:
- Ensure API_KEY and TAVILY_API_KEY are kept secure and not committed to version control.
- The server runs on port 8000 by default; adjust your deployment configuration if needed.
- If you modify code, the --reload flag helps reflect changes automatically during development.
- The endpoints exposed are:
- POST /tavily-search with TavilySearchRequest payload to perform searches via Tavily API.
- POST /tavily-extract with TavilyExtractRequest payload to extract content from URLs via Tavily API.
- When using MCP, ensure your client supplies the required authentication headers or parameters as expected by the server implementation.
- If you run behind a reverse proxy, configure TLS/HTTPS and appropriate headers for API key security.
Related MCP Servers
PPTAgent
An Agentic Framework for Reflective PowerPoint Generation
google_ads_mcp
The Google Ads MCP Server is an implementation of the Model Context Protocol (MCP) that enables Large Language Models (LLMs), such as Gemini, to interact directly with the Google Ads API.
AI-SOC-Agent
Blackhat 2025 presentation and codebase: AI SOC agent & MCP server for automated security investigation, alert triage, and incident response. Integrates with ELK, IRIS, and other platforms.
mcp-document-converter
MCP Document Converter - A powerful MCP tool for converting documents between multiple formats, enabling AI agents to easily transform documents.
ultrasync
MCP server from darvid/ultrasync
fava-trails
🫛👣 FAVA Trails — versioned, auditable memory for AI agents via MCP. Backed by Jujutsu VCS.