At a Glance:
SurfSense is an open-source knowledge management and AI research workspace that provides a self-hostable alternative to Google NotebookLM with support for unlimited sources, 25+ external connectors, a desktop app with system-wide AI assists, and no-code AI automation agents.
Overview:
SurfSense is a self-hosted knowledge base and AI-powered workspace designed to address the source limits, vendor lock-in, and narrow use cases of existing research tools like NotebookLM. It allows users to connect over 25 external data sources, upload files in 50+ formats, and interact with their data through hybrid search, cited chat, and a built-in deliverable studio for generating reports, podcasts, and presentations. The platform supports real-time multiplayer collaboration, configurable LLM and embedding models, and an agentic architecture that enables scheduled or event-triggered automations. A cross-platform desktop app extends SurfSense into any application with quick assist, screenshot assist, and local folder sync capabilities.
Key Decision Points:
Self-hosted or cloud deployment: You can run SurfSense on your own infrastructure with a single Docker command for full data control, or use the managed cloud version at surfsense.com.
Configurable AI backend: You are not locked into a single LLM provider; SurfSense accepts 100+ LLMs via an OpenAI-compatible specification and supports local models through vLLM and Ollama.
Source and size limits: The README states there are no limits on the number of sources or notebooks, and no per-source size cap, in contrast to the documented 200MB limit in NotebookLM.
Agentic automation: SurfSense includes an agent system powered by LangChain Deep Agents that can run on a schedule, trigger when files land in a folder, and write results back to tools like Notion, Slack, and Linear.
Core Features:
25+ external connectors: Sync data from Google Drive, OneDrive, Dropbox, Notion, Slack, GitHub, Jira, Linear, and additional search and productivity services into one searchable index.
Hybrid search with citations: Combine semantic and full-text search over hierarchical indices using reciprocal rank fusion, with responses that include inline source citations.
Deliverable studio: Generate cited research reports, two-host AI podcasts, editable slide presentations, narrated videos, and AI images directly from your document corpus.
Real-time multiplayer: Invite team members to shared search spaces for collaborative AI chat with role-based access control and comment threads.
Desktop application: Launch General Assist from any app, select text for Quick Assist, capture screen regions for contextual questions, and auto-sync a local folder such as an Obsidian vault.
Scheduled and event-driven automations: Build no-code AI workflows by describing them in chat, then run them on a schedule or trigger them when a new document is added to a monitored folder.
Use Cases:
Researchers and analysts who need to work across more than 50 sources and produce structured deliverables like cited reports, podcasts, and presentations.
Developers and technical teams looking for a self-hosted knowledge base that integrates with tools like GitHub, Linear, Slack, and Notion and supports custom LLM configurations.
Users running Obsidian vaults or local document collections who want to make their files searchable and queryable through an AI layer with automatic folder syncing.
Open-Source Alternative Value:
SurfSense is distributed as open-source software that can be self-hosted, allowing users to run the full platform on their own infrastructure without reliance on a single vendor's LLM or storage. The README documents a direct feature comparison with Google NotebookLM, highlighting support for unlimited sources, over 100 configurable LLMs, local model support via Ollama and vLLM, and an agentic architecture—all absent from the compared tool. The project also includes a desktop app and real-time multiplayer, which are not available in the proprietary offering.
