A comprehensive AI tool for generating, styling, and burning professional-grade subtitles and captions into video content locally.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install video-captions
2. Manual Installation
Copy the skill folder to one of these locations
~/.openclaw/skills/ <project>/skills/ Priority: Workspace > Local > Bundled
3. Prompt Installation
Copy this prompt to OpenClaw to install it automatically.
Help me install video-captions using Clawhub. If Clawhub is not installed, install it first (npm i -g clawhub).
Prefer to download?
Get the raw skill files in a ZIP archive.
What is Video Captions?
The Video Captions skill is a powerful utility designed for developers and creators who need high-quality transcription without compromising privacy. By leveraging local engines like Whisper and MLX Whisper, it ensures that your data never leaves your machine unless you explicitly opt for cloud providers. This addition to your library of Openclaw Skills bridges the gap between raw video files and platform-ready content, handling everything from SRT generation to complex ASS styling for social media.
It adheres to professional timing standards, such as Netflix compliance, ensuring that your subtitles are not only accurate but also readable and accessible. Whether you are producing content for YouTube, TikTok, or professional streaming services, this skill automates the tedious aspects of captioning while providing granular control over the final output through integrated ffmpeg processing.
Video Captions Use Cases
- Transcribing video content for YouTube or Vimeo using VTT or SRT formats.
- Creating high-impact, burned-in captions for social media platforms like TikTok and Instagram.
- Generating professional-grade subtitles that meet Netflix-compliant timing and character limit standards.
- Automating multi-speaker identification and diarization for podcasts or interview recordings.
- Translating foreign language video content directly into English subtitles during the transcription process.
How Video Captions Works
- The agent analyzes the input video or audio file to detect the spoken language and select the appropriate transcription engine based on your hardware.
- Transcription is performed locally using Whisper or specialized engines like MLX Whisper for Apple Silicon to ensure 100% data privacy within the Openclaw Skills ecosystem.
- Word-level timestamps are calculated to ensure precise synchronization and enable advanced features like karaoke-style highlighting.
- Subtitles are formatted according to the target platform's requirements, applying specific rules for line breaks, duration, and character limits.
- Optional styling and burn-in are applied via ffmpeg if the user requires embedded captions for social media or specific visual aesthetics.
- A final quality verification check ensures the output meets professional standards before the files are delivered to the local directory.
Video Captions Setup
Ensure you have the required binaries installed on your system. This skill requires ffmpeg and whisper to be available in your PATH.
# Install required binaries on macOS
brew install ffmpeg
pip install openai-whisper
# (Optional) For high-precision word-level timestamps
pip install whisper-timestamped
If you prefer cloud engines, export your API keys as environment variables:
export ASSEMBLYAI_API_KEY='your_api_key_here'
export DEEPGRAM_API_KEY='your_api_key_here'
Video Captions Data Schema & Taxonomy
The skill generates and manages various subtitle formats and temporary processing files as detailed below:
| File Type | Extension | Description |
|---|---|---|
| SubRip | .srt | Universal subtitle format compatible with most video players. |
| WebVTT | .vtt | Preferred format for web-based players like YouTube and HTML5 video. |
| Advanced Substation Alpha | .ass | Used for complex styling, positioning, and social media burn-in. |
| TTML | .ttml | Professional timed text format for broadcast and professional streaming. |
| Metadata | JSON | Internal metadata for tracking transcription accuracy and speaker IDs. |
Video Captions Advanced Features
- Local-First Processing: Runs entirely offline by default, utilizing local hardware acceleration for maximum privacy and speed.
- Professional Compliance: Built-in validation for Netflix-standard line lengths (42 chars), durations, and frame gaps.
- Speaker Diarization: Support for multi-speaker identification using local pyannote or optional cloud APIs for interview processing.
- Advanced Styling Hooks: Direct integration with ffmpeg's styling engine for custom fonts, colors, and dynamic animations.
- VAD Pre-processing: Voice Activity Detection integration to eliminate hallucinations and timing drift during silent video segments.
SKILL.md
Loading
Related Openclaw Skills
Vibe Research
An autonomous AI-led research skill that handles the full cycle from literature review to final synthesis while the human provides high-level vision.
Vibe Marketing
Vibe Marketing is an AI-powered framework for automating content creation and campaign testing by prioritizing creative direction over manual operations.
Vibe Design
Vibe Design leverages AI to generate visual layouts, UI components, and creative assets through descriptive iteration rather than manual pixel-pushing.
Vibe Coding
Vibe Coding is an AI-first development approach where you build software by describing intent rather than writing every line of code.
Video Downloader
A robust AI agent skill for downloading online videos with granular quality controls and secure local storage management.
Video Editing
A sophisticated AI skill designed to automate professional video post-production, from background matting to high-resolution upscaling.
Featured*
