Pay for what you use. Cancel anytime.

How credits work

Instead of charging per file, paid plans give you a monthly pool of credits. Each minute of audio you transcribe spends credits, and the amount depends on the model you choose:

• The fast model and caption fast-path (for URLs that already have subtitles) cost the least.
• Long-form and translation models cost a bit more per minute.
• The highest-quality model (Cohere) costs the most, for when verbatim accuracy matters.

Chat with a transcript and semantic search do not draw down credits. Generated outputs — including notes, chapters, flashcards, action items, and Work Products — use at least 1 credit per generation. The charge increases by 1 credit for each started 100,000 characters of transcript context. The app checks the charge against your remaining balance before generation.

Published usage limits

• Transcription count: Free includes 3 lifetime transcriptions; Go includes 50, Pro 200, and Max 500 per monthly quota period.
• Generated outputs: up to 50 notes, chapters, flashcards, action items, or Work Products in a rolling 24-hour period on Pro and Max.
• Transcript chat: up to 200 user messages in a rolling 24-hour period on paid plans.
• Rolling 24-hour audio ceiling: Free 60 minutes, Go 8 hours, Pro 20 hours, Max 40 hours.
• Simultaneous transcription jobs: Free 1, Go 3, Pro 6, Max 10.
• Short-term request limits protect the service from bursts. Max has higher chat and search request limits, but no separate priority queue.

Which plan is right for you?

• Free — trying it out, short clips, occasional use.
• Go — students and creators who transcribe lectures, videos, and podcasts regularly.
• Pro — professionals and researchers who need every model, speaker labels, and search across many recordings.
• Max — power users with very long recordings and high monthly volume.

Not sure? See how it works for lectures, YouTube videos, or podcasts.