Transcribe your meetings without sending audio anywhere
Fly Scribe captures your microphone and the audio of a browser tab and turns them into text right in your browser with Whisper. No servers, no accounts.
StartTranscription
Transcript
The transcript will appear here. Press “Start” to begin.
How it works
- 1
Choose your sources
Enable your microphone, the audio of a tab (Meet, Teams, YouTube…), or both.
- 2
The model runs on your device
Whisper is downloaded once (~80 MB) and runs in your browser. Audio never leaves your computer.
- 3
Read and export
Text appears labeled and timestamped. Download it as Markdown or copy it.
Frequently asked questions
Does my audio leave my computer?
No. Transcription runs locally in your browser. The only download is the Whisper model, which is cached afterwards.
Why do I only get tab audio on macOS?
Browsers can only capture tab audio on macOS. Desktop apps need third-party software. Use Teams or Meet in a browser tab instead.
How accurate is it?
The “base” model is a balance of size and speed. It struggles more with Chinese, Japanese and Russian, noisy audio and overlapping voices. Fixing the audio language is more accurate than auto-detect.
Why does the text arrive late?
Whisper is not a streaming model, so audio is transcribed in chunks of about 6 seconds.
Do I need to tell people I am recording?
Yes. You are responsible for informing participants and complying with local regulations.