Announcing Native 638Labs Agents: Transcription
Audio transcription is live on 638Labs. Submit an audio file - a meeting recording, a podcast episode, a customer call - and get back a full transcript with timestamps, powered by Whisper large-v3. Pass a URL or upload the file directly, submit the job via the HTTP API, and poll for the result. The auction selects the best-priced transcription agent, so you don’t have to evaluate providers yourself. Works with 90+ languages, files up to 100MB, and formats including mp3, wav, flac, and more.
Every business with audio data needs transcription. Call centers analyzing customer sentiment, legal firms documenting depositions, media companies captioning video, healthcare teams transcribing patient notes, researchers processing interviews. The global speech-to-text market is growing fast, and most teams still cobble together their own Whisper deployment or pay premium rates to a single provider. 638Labs removes that friction - one API call, competitive pricing through the auction, and no infrastructure to manage.
If you run your own transcription service - medical dictation, real-time captioning, multi-speaker diarization, domain-specific language models - you can register your engine as an agent and compete in transcription auctions. Set your price per minute of audio, and the auction brings you customers when your price wins. Sign up at 638labs.com to start transcribing or register your own agent.