Open-Notebook

An independent consultant handling confidential client archives cannot upload proprietary PDF documents into a public cloud assistant without violating privacy agreements. Open-Notebook replicates the document-grounded research workflow on private infrastructure, letting researchers query files and generate multi-speaker audio summaries through self-hosted software.

Why Open-Notebook instead of NotebookLM

NotebookLM delivers a zero-maintenance service inside Google's cloud. Users sign in with a Google account, upload files, and immediately query notes while Google manages embeddings, compute, and storage behind the scenes. Open-Notebook shifts the operational burden entirely to the user. Running as a Docker Compose stack pairing the application with SurrealDB, it requires configuring storage volumes, network ports, and encryption keys. In return, organizations gain data sovereignty: uploaded files, stored embeddings, and query histories remain on local infrastructure.

Operational differences also define how language models run. NotebookLM restricts interactions to Google Gemini models, enforcing quota ceilings on document uploads and audio generation lengths. Open-Notebook decouples the user interface from the underlying model provider, allowing operators to connect API keys from OpenAI, Anthropic, or Groq, or link directly to local Ollama instances. This flexibility requires ongoing maintenance: administrators must manage API credentials, track token expenses, or supply dedicated GPU memory when running local models.

What it does well

• Multi-speaker audio generation: creates custom podcasts with up to four distinct voices, turning written research into engaging spoken dialogue.
• Broad document intake: ingests PDF files, text notes, web links, and YouTube video transcripts into a single workspace without manual text conversion.
• Flexible model routing: lets users assign different AI providers for chat, text embeddings, and audio generation to balance cost and analytical quality.
• Local offline execution: pairs with Ollama to run summaries and document queries on a desktop workstation without internet access.
• Isolated notebook workspaces: separates research materials, uploaded sources, and conversation logs into distinct projects, preventing unrelated notes from contaminating search queries.
• Automated study aids: generates structured outlines, executive summaries, and study notes directly from uploaded sources to speed up reading workflows.