💡 GGUF models run entirely in your browser via llama.cpp. First load may take a few minutes (cached after).

Document Q&A

Upload a PDF or text file and ask questions answered only from that document — fully in your browser, nothing uploaded to a server.

Initializing Agent

Starting up…

⏳ Loading Python runtime
⏳ Loading packages
⏳ Setting up tools
⏳ Preparing agent