NoteCast — AI tab-audio note-taker

About

NoteCast turns anything playing in a browser tab — lectures, meetings, videos, podcasts — into a live transcript and structured notes (summary, key points, action items), on your own machine.

- Your models, your machine: audio is always transcribed on-device with Whisper (WebGPU/WASM), it never leaves your computer. Notes are generated by Ollama, running wherever you want it: locally by default (http://localhost:11434), or a remote/self-hosted Ollama server you point it at. No third-party cloud, no account, no vendor API keys.

- Live notes: the transcript grows every ~20–30 seconds and notes regenerate as you watch.
- YouTube-aware: every transcript line links to the exact moment in the video; navigating to a new video mid-recording splits the session automatically.
- Keyboard shortcut (Ctrl/Cmd+Shift+S), session history, Markdown and plain-text export.

Requires: Chrome/Brave 116+, and Ollama with a model pulled — running locally (default) or on a remote host you set in Settings (e.g. ollama pull llama3.1). First use downloads Whisper weights from Hugging Face (cached afterwards). If Ollama needs a one-time config tweak to accept the extension, NoteCast detects it and shows the exact command to run.

Privacy: your audio is transcribed on-device and never leaves your machine. Transcripts and notes are stored locally; if you choose a remote Ollama host, only the transcript text is sent to that server, one you run yourself. No analytics, no telemetry, nothing sent to the developer. See the privacy policy for details.