NotebookLM without the Google-shaped cage

What it does

SurfSense ingests documents from 27-plus external sources—Google Drive, Slack, Notion, GitHub, Gmail, Obsidian, and more—into searchable knowledge bases called SearchSpaces. Teams can chat with their data, generate cited reports, podcasts, slide decks, and videos, or automate workflows that write results back to external tools. A companion desktop app adds global shortcuts, screenshot analysis, and local folder sync so your private knowledge base follows you across applications.

The interesting bit

The project treats vendor independence as a first-class feature: it supports over 100 LLMs via the OpenAI spec and LiteLLM, more than 6,000 embedding models, and local inference through Ollama or vLLM, all wrapped in a Docker-based deployment that keeps data on your infrastructure. Real-time multiplayer chat with comments and RBAC is something the original tool simply does not offer.

Key highlights

• Unlimited sources and notebooks with no per-file word or size caps, according to the README.
• Hybrid semantic and full-text search with hierarchical indices and reciprocal rank fusion.
• Agentic architecture using LangChain Deep Agents with planning, subagents, and file-system access.
• Desktop app and browser extension for global shortcuts, screenshot analysis, Obsidian sync, and saving auth-protected pages.
• AI automations and file sorting that run on schedules or trigger the moment a document lands in a folder.

Caveats

• The README explicitly states the project is not yet production-ready and is actively being developed.
• Some generative features—video and podcast creation—are acknowledged to lag behind NotebookLM’s quality.

Verdict

Teams and privacy-conscious users who need multiplayer research notebooks and self-hosted LLMs should watch this closely. If you require a polished, battle-tested SaaS today, the authors warn it is not quite there yet.