Upload spoken MP3 audio, review the timestamped transcript, then export TXT, SRT or Markdown. Check names and numbers before reuse.
Free: up to 100MB and 30min per file. Transcription uses credits based on duration. Source files are temporary; keep your own copy.
1 min × 3 credits/min = 3 credits
Play the file and find a passage with clear speech. MP3 compression, music and overlapping voices can obscure words. The example below pairs an authored reference transcript with synthetic speech; it is a format demo, not an AI accuracy test.
Replay the matching timestamp and correct names, numbers and specialist terms. A silent file will not produce useful text. If the upload fails, check the file limit and connection before starting a new transcription.
TXT and Markdown work well for reading and editing; SRT keeps subtitle timing. VTT, STL, XML, FCPXML and EDL require Creator Plus or Studio. Check timings in your editor before publishing.
AI transcription costs 3 credits per rounded-up minute; a summary costs 20; translation costs 5 per rounded-up minute for each target language. These are separate tasks sharing one balance. New accounts receive 50 Starter Credits once, not a monthly allowance.
A self-authored English sample with synthetic speech and a matching reference transcript. This is a format demonstration, not an AI accuracy benchmark or a customer recording.
English sample · 15 seconds
Download sample MP3Playing this example and downloading its reference files use no credits.
A useful research note keeps the claim connected to its source.
First, save the exact words with a time reference.
Next, write a short summary in your own words.
Finally, listen again to check names and numbers before you share the note.
Your own transcription requires sign-in and credits. File and duration limits apply; check the estimate before processing.
Workflow
Upload spoken MP3 audio, review the timestamped transcript, then export TXT, SRT or Markdown. Check names and numbers before reuse.
Check the duration and estimated credits before confirming. Follow the current job status instead of starting duplicate jobs.
Listen to uncertain passages again. Correct names, numbers and quotations, then choose readable text or timed subtitles for your next task.
