mpv-sub-fastwhisper

Project Overview

mpv-sub-fastwhisper is a subtitle-generation script integrated into mpv that uses faster-whisper to transcribe video audio and output SRT subtitles. After generating subtitles, the script also tries to translate them through the GPT API, making it quick to get subtitled playback. It suits users who want to generate subtitles right during playback.

One step during playback and mpv writes the subtitles for you.

Key Features

• Speech transcription: uses faster-whisper to transcribe video audio into SRT subtitles
• Automatic translation: tries to translate the generated subtitles via the GPT API
• Script integration: runs as an mpv script, generating subtitles during playback

Target Audience

• Users watching videos without subtitles
• Users of mpv who need transcribed subtitles
• Users with local speech-to-text and translation needs

Quick Start

1. Install faster-whisper and FFmpeg
2. Place the script in mpv's scripts directory
3. Fill in the GPT API configuration in the script or its options
4. Open a video in mpv and trigger subtitle generation
5. The generated subtitle file is saved alongside the video

Core Advantages

• Local transcription: transcription happens on your machine, with no video upload
• Translation included: it tries to translate after transcribing, saving a second pass
• Fits playback: it generates inside mpv, so you wait for subtitles while watching

System Requirements

Item	Requirement
OS	Any platform that runs mpv
Dependencies	faster-whisper, FFmpeg
Runtime	mpv
Optional	The GPT API (subtitle translation)

Notes

• GPT API settings must be filled in yourself in the script or its options
• Transcription speed and accuracy depend on your machine and the model used
• Using online translation incurs the corresponding API costs