A self-hosted AI research notebook that organizes multi-modal content and provides chat, search, and podcast generation. It supports 18+ AI providers and can be deployed via Docker without coding for the basic setup.
Project overview
Provides a self-hostable alternative to cloud-only research notebooks, allowing users to choose among 18+ AI providers including local model options.
Project type
MCP · RAG
Use cases
Knowledge Q&A
Deployment
Refer to project documentation
License
MIT
Best for
Researchers and general users needing a self-hosted research notebook with multi-modal content ingestion and multi-model AI provider support.
Key capabilities
Organize and ingest multi-modal content including PDFs, videos, audio, web pages, and Office docs.
AI conversations powered by your research materials.
Full-text and vector search across all ingested content.
Advanced multi-speaker podcast generation with custom Episode Profiles. Supports 1-4 speakers with custom profiles.
Customizable actions to summarize and extract insights from content.
Support for 18+ AI providers for LLMs, embeddings, speech-to-text, and text-to-speech.
Full programmatic access for custom integrations.
Provides basic references for answers, with plans to improve to comprehensive citations.
Limitations and risks
Citation features currently provide basic references compared to comprehensive citations with sources, though improvements are planned.
Telemetry practices and minimum hardware/GPU requirements are not documented.
Getting started
Setup difficulty is documented as easy because Docker compose is the default path and no code is required for basic setup. The documented first success path is to download docker-compose.yml, edit the encryption key, run docker compose up -d, and open http://localhost:8502.
Alternatives and comparisons
Provides a single knowledge base application for individuals and agents that turns scattered documents into searchable, connected, and traceable knowledge using an original retrieval architecture.
An autonomous, self-maintaining personal Wikipedia built by AI that transforms documents, notes, and web clippings into a permanent knowledge base with cross-linking, graph viewing, and citations.
GitHub project description: An Open Source implementation of Notebook LM with more flexibility and features
README: Open Notebook empowers you to: - 🔒 **Control your data** - Keep your research private and secure - 🤖 **Choose your AI models** - Support for 18+ providers including OpenAI, Anthro…
README: - 📚 **Universal Content Support**: PDFs, videos, audio, web pages, Office docs, and more - 🤖 **Multi-Model AI Support**: 18+ providers including OpenAI, Anthropic, Ollama, Google,…
README: | **Privacy & Control** | Self-hosted, your data | Google cloud only | Complete data sovereignty |
README: - **🔧 Content Transformations**: Powerful customizable actions to summarize and extract insights - **🌐 Comprehensive REST API**: Full programmatic access for custom integrations