open-notebook

A self-hosted AI notebook centralizes literature, PDFs, web pages, and media into linked sources for full-text and vector search. You organize content into notebooks, generate AI-assisted notes, and engage in context-aware chats that cite sources during research discussions.

Semantic search across local content uses vector embeddings to find conceptually matching text within your ingested research materials. It works alongside keyword search to retrieve relevant information from notebooks, sources, and notes without sending data to external services.

Yes, you can ingest PDFs, web pages, and media into the notebook via URLs, uploads, or text notes. The self-hosted deployment provides REST API access, allowing automated document ingestion and management of research sources programmatically.

Docker-based deployment supports multiple AI providers for end-to-end research workflows. You configure Docker Compose to launch the local research workspace, which then connects to your chosen AI providers for context-aware chat, note generation, and content transformations.

Yes, you can create narrated summaries or podcasts directly from notebook content. The system applies transformations to extract methods, entities, and insights, then generates audio output to summarize your research materials.

Self-hosted AI notebooks require local infrastructure management and Docker setup for document ingestion. You must provision sufficient storage for vector search indexes and manage API keys for external AI providers, as processing relies on your configured services rather than managed cloud solutions.