Plain-AyI
← All projects
Speech & audio

WhisperX

Fast speech-to-text that gives a time for every word and labels who is speaking.

Research-backed

What it does

WhisperX builds on Whisper, OpenAI's speech-recognition model, to make transcripts faster and more precise. Its makers say it can transcribe many times faster than real time on a capable graphics card.

Plain Whisper gives times for whole sentences, which can be off by several seconds. WhisperX lines up the text with the audio so each word gets its own timestamp, which helps for subtitles and editing.

It can also tell different speakers apart and label them, a step called speaker diarization. It filters out silent parts first, which its makers say reduces made-up text.

Who it's for

Journalists, researchers, podcasters and developers who need detailed transcripts or subtitles.

Can I use it without coding?

With some setup. Installed with Python and run from the command line. A graphics card is recommended for speed. New to this? Start here.

Can I run it?

Cost
FreeNothing to pay or sign up for.
Runs on
Mac, Windows, Linux
Computer
Works best on a powerful computer or one with a graphics card (GPU)
Privacy
Runs on your own device: your data stays with you
API key
Not needed

Try it

  1. Install WhisperX with Python's package installer.
  2. Run it on an audio file to get a transcript with a time for every word.

Using Claude Code, Cursor or another AI coding assistant? Paste the prompt and it will walk you through the setup.New to this?

Before you install

  • Only download from the official GitHub page. Copies with similar names exist for many popular projects.
  • License (BSD-2-Clause): Permissive: free to use, change and share, including for business.
  • More on staying safe

Where people are talking about it