This hosted release supports English speech only. Other spoken languages are not supported yet.
Terms of service
Effective October 15, 2026. Using the service means accepting these terms and the privacy policy.
The service
WhisperX is an independent hosted interface for transcription technology. It provides file upload, asynchronous processing, transcript editing, and downloads. It is not affiliated with the WhisperX authors, OpenAI, or pyannote. The open source software remains available under its own terms.
Accounts and uploads
Use accurate account information and protect your login. Real processing requires an account. Upload only media you have the right to process and share with the service’s infrastructure providers, with consent where required. Do not use the service for unlawful surveillance, unauthorized personal data collection, copyright infringement, or abusive content. We may limit or suspend misuse.
Minutes and prices
The initial verified-account trial includes 5 minutes once and expires 30 days after grant. Starter provides 120 minutes for $9 USD; Creator provides 600 minutes for $29 USD. Purchased minutes expire 12 months after purchase. These are one time purchases with no automatic renewal. Any applicable tax and final payment total are shown at checkout.
Processing and charges
All transcription modes use the same duration based pricing. We round confirmed media duration up to the next whole second rather than rounding every task to a whole minute. A task reserves its required balance before processing and settles after successful delivery. A failed task releases the reservation. Editing and exporting an existing result do not use more minutes. Starting a new transcription is a separate charge.
Output and review
Model output may contain incorrect text, incomplete timing, or inaccurate speaker labels. Check the recording before relying on the result. Text edits do not automatically create new aligned word times. The service does not establish a speaker’s identity and should not be treated as an authoritative evidentiary record or professional advice.
Availability and retention
The service can queue tasks, encounter provider outages, or reject unsupported media. Input and output retention follow the privacy policy and expiry shown on the task. Download what you need before expiry. File deletion after successful delivery does not automatically return consumed minutes. We may change the service and will present the current configuration and charge before a new task.
Payment questions and refunds
A processing failure releases reserved minutes; it is separate from a refund of a purchased pack. For a mistaken purchase, duplicate charge, or service issue, contact support@whisperx.app with the payment reference. Refund requests are reviewed according to the facts and applicable consumer rights. We do not promise an automatic cash refund for every inaccurate transcription.
Responsibility and contact
You are responsible for your recording rights and use of output. To the extent permitted by applicable law, the service is supplied without a guarantee of uninterrupted availability or suitability for a particular purpose. Nothing here removes mandatory consumer rights. Contact support@whisperx.app for service, legal, or rights questions.
Read the upstream WhisperX project · Contact support