Official documentation

WhisperX: README (GitHub, m-bain/whisperX)

m-bain/whisperX project, GitHub (m-bain/whisperX), 2026

Read on October 17, 2026

What it is: The README of WhisperX, an open-source tool built on Whisper and faster-whisper that adds batched inference, word-level timestamps and speaker diarization. It states its speed figure and the reason for its timestamp alignment.
How much weight it carries: Maintained by the project's authors, so the speed figure is their own measurement with a stated model and setup. The README describes the problem it fixes, which is independent of the speed claim.

Facts taken from this source

WhisperX reports 70 times real-time transcription with the large-v2 model, using batched inference.
“This repository provides fast automatic speech recognition (70x realtime with large-v2) with word-level timestamps and speaker diarization.” (Introduction)
Whisper's own timestamps are per utterance rather than per word, and can be inaccurate by several seconds.
“the corresponding timestamps are at the utterance-level, not per word, and can be inaccurate by several seconds” (Introduction)
WhisperX, an open-source project that runs Whisper-based recognition, adds word-level timestamps and speaker diarization to its output.
“This repository provides fast automatic speech recognition (70x realtime with large-v2) with word-level timestamps and speaker diarization.” (Introduction)
The WhisperX README lists the faster-whisper project as the backend it uses for batched inference.
“faster-whisper](https://github.com/guillaumekln/faster-whisper) backend” (Features)

Official documentation

m-bain/whisperX project, GitHub (m-bain/whisperX), 2026

Read on October 17, 2026

What it is: The README of WhisperX, an open-source tool built on Whisper and faster-whisper that adds batched inference, word-level timestamps and speaker diarization. It states its speed figure and the reason for its timestamp alignment.
How much weight it carries: Maintained by the project's authors, so the speed figure is their own measurement with a stated model and setup. The README describes the problem it fixes, which is independent of the speed claim.

WhisperX reports 70 times real-time transcription with the large-v2 model, using batched inference.
“This repository provides fast automatic speech recognition (70x realtime with large-v2) with word-level timestamps and speaker diarization.” (Introduction)
Whisper's own timestamps are per utterance rather than per word, and can be inaccurate by several seconds.
“the corresponding timestamps are at the utterance-level, not per word, and can be inaccurate by several seconds” (Introduction)
WhisperX, an open-source project that runs Whisper-based recognition, adds word-level timestamps and speaker diarization to its output.
“This repository provides fast automatic speech recognition (70x realtime with large-v2) with word-level timestamps and speaker diarization.” (Introduction)
The WhisperX README lists the faster-whisper project as the backend it uses for batched inference.
“faster-whisper](https://github.com/guillaumekln/faster-whisper) backend” (Features)