SurfSense

以下为收录时的原文，语言与页面不同，未做翻译。

SurfSense is an open-source team collaboration tool built as an alternative to NotebookLM, connecting LLMs to internal knowledge sources for real-time chat, research, and workflow automation with cited answers.

官网要点

• Reddit API
• Instagram API
• Google Maps API
• Indeed API

官方常见问题

What is open web research?

Open web research is gathering and analyzing live public information from across the web: search results, community discussions, reviews, videos, and any page. Unlike asking a chatbot that reasons over a stale index, it works from what the web says right now. SurfSense automates it: AI agents collect the live data and turn it into cited briefs and alerts.

What is an MCP server?

An MCP server exposes tools and data to AI agents through the Model Context Protocol, an open standard adopted by Claude, Cursor, and most agent frameworks. Add the SurfSense MCP server and your agents can call every connector, such as reddit.scrape or google_search.scrape, as native tools.

How is SurfSense different from a web scraping API?

A web scraping API returns raw data and leaves the intelligence to you. SurfSense pairs platform-native connectors with an agent harness: retries, structured output, credit metering, and an MCP server, so your agents go from a question to a brief without you building the plumbing in between.

Can I use the connector APIs directly in my own app?

Yes. Every platform connector is a typed REST endpoint you can call from any language with your SurfSense API key, no agent required. Send a POST request with your query and get structured JSON back. Each connector page has copy-paste examples in cURL, Python, JavaScript, Go, and more.

Can I self-host SurfSense?

Yes. SurfSense is open source and self-hostable, so you can run the entire platform on your own infrastructure and keep sensitive research in-house. Use the cloud version to start in minutes, or deploy from the GitHub repository when you need full control.