WhisperX: a place to run WhisperX
You have already chosen the technology. We help you get to the result.
Built for people who want WhisperX
WhisperX serves people who know they want WhisperX but would rather skip Python environments, accelerator configuration, model downloads, and ongoing maintenance. Upload your own recording, review the transcript, and download output you can use in your next workflow.
What the hosted service adds
The web application handles accounts, private uploads, task tracking, minute balances, playback, text corrections, speaker labels, and export formats. Speech processing happens in a separate hosted inference environment. The browser is an editor and upload interface; it does not run the speech model locally.
Independent and unofficial
WhisperX is independently operated. It is not the upstream WhisperX project and is not affiliated with or endorsed by its authors, OpenAI, or pyannote. The project’s maintainers do not provide customer support for this website. Please direct service questions to support@whisperx.app.
A clear boundary around the output
Automatic speech recognition, alignment, and speaker grouping can make mistakes. We keep original output distinct from your edits and do not promise a universal accuracy rate or processing speed. Check important statements against the recording before publication or use in a consequential decision.
Support and open source
For account, upload, task, and payment questions, contact support@whisperx.app and include the task identifier when available. For the open source project, read m-bain/whisperX on GitHub. Please do not include private recordings or credentials in a public issue.
Read the upstream WhisperX project · Contact support