{"type":"blog_post","title":"SurfSense MCP Server: Open Web Research for Your Agents","description":"SurfSense is an open-source NotebookLM alternative designed for open web research with live data. It provides a suite of connectors as native tools, accessible to your agents via its MCP server. Developers can integrate SurfSense with platforms like Claude or Cursor to empower AI agents with real-time web research capabilities.","content":"# SurfSense MCP Server: Open Web Research for Your Agents\n\nSurfSense offers an open-source NotebookLM alternative, purpose-built for conducting open web research with live data. It provides a unified platform, API, or MCP server to access its research capabilities. For developers building AI agents, SurfSense's MCP server integration is a direct path to equipping those agents with real-time web data access.\n\n## Wiring SurfSense into Your Agent Framework\n\nIntegrating SurfSense with your agent framework, whether it's Claude, Cursor, or a custom build, involves configuring the MCP server endpoint and authentication. The core of this integration is the `mcpServers` configuration block.\n\nTo connect your agent to the hosted SurfSense MCP server, you'll use the following JSON configuration:\n\n```json\n{\n  \"mcpServers\": {\n    \"surfsense\": {\n      \"url\": \"https://mcp.surfsense.com/mcp\",\n      \"headers\": { \"Authorization\": \"Bearer ${SURFSENSE_API_KEY}\" }\n    }\n  }\n}\n```\n\nThis configuration points your agent to the official SurfSense MCP server URL and includes a placeholder for your `SURFSENSE_API_KEY` for authorization. Once configured, your agent can call every connector exposed by SurfSense as if it were a native tool.\n\nFor those preferring to host the server themselves, SurfSense also supports running the server locally from the `surfsense_mcp` module. This provides flexibility for development and custom deployments.\n\n## Powering Agents with Web Connectors\n\nThe primary benefit of integrating SurfSense via MCP is providing your agents with a comprehensive set of web research tools. The SurfSense MCP server exposes its full list of connectors as callable tools, enabling agents to perform actions like:\n\n*   **Live data retrieval:** Agents can access current information directly from the open web, moving beyond static training data.\n*   **Structured research:** The underlying SurfSense platform is designed for organized web research, which translates into more effective tool use by agents.\n\nThis capability makes SurfSense particularly useful for agents that need to stay current with information, verify facts, or gather data on dynamic topics.\n\n## References\n- [SurfSense on GitHub](https://github.com/MODSetter/SurfSense)\n- [Model Context Protocol Documentation](https://modelcontextprotocol.io/introduction)\n- [SurfSense on model-context-protocol.com](https://model-context-protocol.com/servers/)","keywords":["surfsense","mcp-server","modsetter-surfsense","open-web-research","ai-agents"],"published_at":"2026-07-22T12:00:33.63+00:00","related_repository":{"slug":"surfsense","type":"Server","url":"https://model-context-protocol.com/servers/surfsense"},"source_url":"https://model-context-protocol.com/blog/surfsense-mcp-server-open-web-research-for-your-agents-mcp-server-guide"}