Back to Skills

video-translation

NoizAI
Updated 2 days ago
5 views
502
74
502
View on GitHub
Othergeneral

About

This skill translates and dubs videos by downloading the source, extracting subtitles, translating the text, generating new audio with TTS, and replacing the original audio track. It's designed for use cases like localizing YouTube videos when a user requests a language change. Key capabilities include handling video downloads, subtitle processing, and audio synthesis while preserving the original video.

Quick Install

Claude Code

Recommended
Primary
npx skills add NoizAI/skills -a claude-code
Plugin CommandAlternative
/plugin add https://github.com/NoizAI/skills
Git CloneAlternative
git clone https://github.com/NoizAI/skills.git ~/.claude/skills/video-translation

Copy and paste this command in Claude Code to install this skill

Documentation

Video Translation

Translate a video's speech into another language, using TTS to generate the dubbed audio and replacing the original audio track.

Triggers

  • translate this video
  • dub this video to English
  • 把视频从 X 语译成 Y 语
  • 视频翻译

Use Cases

  • The user wants to watch a foreign language YouTube video but prefers to hear it in their native language.
  • The user provides a video link and explicitly requests changing the audio language.

Workflow

When the user asks to translate a video:

  1. 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
    
  2. 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.

  3. 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
    
  4. 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
    
  5. Present the Result: Return the final_video.mp4 file path to the user.

Inputs

  • Required inputs:
    • VIDEO_URL: The URL of the video to translate.
    • target_language: The language to translate the audio to.
  • Optional inputs:
    • 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.

Outputs

  • Success: Path to the final video file with replaced audio.
  • Failure: Clear error message specifying whether download, TTS, or audio replacement failed.

Requirements

  • Dependencies (other skills)
    • youtube-downloader (crazynomad/skills) — SKILL.md
      Install: clone or copy the skills/youtube-downloader directory from crazynomad/skills into your skills/ folder so that skills/youtube-downloader/scripts/download_video.py is available.
    • tts (NoizAI/skills) — SKILL.md
      If not already in this repo: clone or copy the 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.

Limitations

  • The source video must have subtitles (or auto-generated subtitles) available on the platform for the source language.
  • Very long videos may take a significant amount of time to translate and dub.

GitHub Repository

NoizAI/skills
Path: skills/video-translation
0

Related Skills

llamaguard

Other

LlamaGuard is Meta's 7-8B parameter model for moderating LLM inputs and outputs across six safety categories like violence and hate speech. It offers 94-95% accuracy and can be deployed using vLLM, Hugging Face, or Amazon SageMaker. Use this skill to easily integrate content filtering and safety guardrails into your AI applications.

View skill

cost-optimization

Other

This Claude Skill helps developers optimize cloud costs through resource rightsizing, tagging strategies, and spending analysis. It provides a framework for reducing cloud expenses and implementing cost governance across AWS, Azure, and GCP. Use it when you need to analyze infrastructure costs, right-size resources, or meet budget constraints.

View skill

quantizing-models-bitsandbytes

Other

This skill quantizes LLMs to 8-bit or 4-bit precision using bitsandbytes, achieving 50-75% memory reduction with minimal accuracy loss. It's ideal for running larger models on limited GPU memory or accelerating inference, supporting formats like INT8, NF4, and FP4. The skill integrates with HuggingFace Transformers and enables QLoRA training and 8-bit optimizers.

View skill

dispatching-parallel-agents

Other

This Claude Skill dispatches multiple agents to investigate and fix 3+ independent problems concurrently. It is designed for scenarios involving unrelated failures that can be resolved without shared state or dependencies. The core capability is parallel problem-solving, assigning one agent per independent problem domain to maximize efficiency.

View skill