Translate and dub videos from one language to another, replacing the original audio with TTS while keeping the video intact.
Translate a video's speech into another language, using TTS to generate the dubbed audio and replacing the original audio track.
When the user asks to translate a video:
Download Video & Subtitles:
Use the youtube-downloader skill to download the video and its subtitles as SRT. Make sure you specify the source language to fetch the correct subtitle.
python path/to/youtube-downloader/scripts/download_video.py "VIDEO_URL" --subtitles --sub-lang <source_lang_code> -o /tmp/video-translation
Translate Subtitles:
Read the downloaded .srt file. Translate its contents sentence by sentence into the target language using the following fixed prompt. Keep the exact same SRT index and timestamp format!
Translation Prompt:
Translate the following subtitle text from <Source Language> to <Target Language>. Provide ONLY the translated text. Do not explain, do not add notes, do not add index numbers. The translation must be colloquial, natural-sounding, and suitable for video dubbing.
Save the translated text into a new file translated.srt.
Generate Dubbed Audio:
Use the tts skill to render the timeline-accurate audio from the translated SRT. The Noiz backend automatically aligns the duration of each sentence to the original video's subtitle timestamps.
To ensure the cloned voice matches the original speaker's exact tone and emotion for each sentence, pass the original video file to --ref-audio-track. The TTS engine will automatically slice the original audio at each subtitle's exact timestamp and use it as the reference for that specific segment.
Create a basic voice_map.json:
{
"default": {
"target_lang": "<target_lang_code>"
}
}
Render the timeline-accurate audio:
bash skills/tts/scripts/tts.sh render --srt translated.srt --voice-map voice_map.json --backend noiz --auto-emotion --ref-audio-track original_video.mp4 -o dubbed.wav
Replace Audio in Video:
Use the replace_audio.sh script to merge the original video with the new dubbed audio. To keep the original video's non-speech audio background outside of translated segments, pass the --srt file.
bash skills/video-translation/scripts/replace_audio.sh --video original_video.mp4 --audio dubbed.wav --output final_video.mp4 --srt translated.srt
Present the Result:
Return the final_video.mp4 file path to the user.
VIDEO_URL: The URL of the video to translate.target_language: The language to translate the audio to.source_language: The language of the original video (if not auto-detected or specified).reference_audio: Specific audio file/URL to use for voice cloning instead of the dynamic original video track.skills/youtube-downloader directory from crazynomad/skills into your skills/ folder so that skills/youtube-downloader/scripts/download_video.py is available.skills/tts directory from NoizAI/skills into your skills/ folder. Ensure skills/tts/scripts/tts.sh and related scripts are present.NOIZ_API_KEY configured for the Noiz backend. If it is not set, first guide the user to get an API key from https://developers.noiz.ai/api-keys. After the user provides the key, ask whether they want to persist it; if they agree, either write/update NOIZ_API_KEY=... in the project's .env file or run bash skills/tts/scripts/tts.sh config --set-api-key YOUR_KEY to store it.ffmpeg installed.