Transcription Service
Submit a public audio URL and get back a speaker-labelled transcript with [hh:mm:ss] timestamps — the whole recording transcribed in one shot, no upload required.
New accounts are created automatically on first sign-in.
One-shot transcription
Paste a link to an mp3, wav, m4a, or opus recording. The service sends the URL straight to the model and transcribes the full recording in a single call.
Timestamped speakers
Transcript lines follow the [hh:mm:ss] Speaker: text format, so you can jump to any moment and see who said what.
Usage tracking
Every job records its token usage. The usage dashboard shows job count, total derived audio duration, token breakdowns, and an estimated cost.
API keys
Create API keys for programmatic access. Secrets are shown once at creation and stored only as a verifiable hash.
How it works
- Sign in with Google. No separate account or password.
- Paste a public audio URL. The URL is checked for reachability and a supported audio format before the job is queued.
- Watch it finish. Job status updates in place — queued, transcribing, done — and the full transcript opens when it completes.
Supported formats
mp3, wav, m4a, and opus. The audio must be publicly reachable over http or https; the service hands the URL to the model and never stores the audio itself.