About Notate
Notate is an open source desktop application that acts as a private AI research companion. Users upload documents, run analysis on content and get answers drawn from their own material, with the option to keep all data on their own machine.
Documents, YouTube videos and web pages can be analyzed, and an advanced web crawler can pull in site content. Knowledge is stored in a ChromaDB vector database and searched with semantic retrieval across the whole library. For models, Notate connects to OpenAI, Anthropic, Gemini and XAI, and it can run open source models fully offline through Ollama with local embeddings. Providers can be mixed or the app can run entirely on local models.
It is aimed at researchers, students and developers who want a chat and retrieval tool over their own files without sending everything to a cloud service. The project is built with TypeScript, React, Python, FastAPI and ChromaDB, and released under the Apache License 2.0, so it can be modified and extended. Native apps are offered for macOS, Windows and Linux. System requirements depend on mode. External API mode needs Python 3.12, 4 GB of RAM and 2 GB of free disk space. Local model mode needs 16 GB of RAM, with 32 GB recommended, 10 GB or more of disk for models and files, a modern multi core CPU and a GPU with 8 GB or more of VRAM recommended.
The software is free to download. The site publishes no paid plans, and any model provider charges would apply separately.