An efficient tool to transcribe local video or audio files into timestamped SRT subtitles and plain text documents.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install maple-video2txt
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 maple-video2txt 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 Video2txt: AI-Powered Media Transcription?
Video2txt is a high-performance utility designed for Openclaw Skills users who require accurate, local transcription of media files. By utilizing the faster-whisper engine, it processes audio and video content directly on your machine, ensuring data privacy and reducing reliance on external APIs. This skill is optimized for developers and content creators who need to transform speech into structured text formats quickly.
Whether you are handling meetings, lectures, or creative content, Video2txt provides a streamlined workflow for generating subtitles and transcripts. It features built-in support for multiple languages, with a specific focus on Chinese language processing, including automatic conversion from traditional to simplified characters for improved compatibility and readability.
Video2txt: AI-Powered Media Transcription Use Cases
- Creating SRT subtitle files for video tutorials or vlogs.
- Transcribing business meetings and interviews into text for documentation.
- Extracting text from audio lectures to create study notes within the Openclaw Skills environment.
- Converting old video archives into searchable text databases.
- Generating transcripts for podcasts to improve accessibility and SEO.
How Video2txt: AI-Powered Media Transcription Works
- The user provides a path to a local media file (video or audio) via the command line.
- The skill initializes the faster-whisper engine and loads the specified AI model.
- Audio is extracted from the source file using ffprobe/ffmpeg and passed to the transcription model.
- The AI model performs speech-to-text conversion, identifying timestamps and segments.
- If the content is in Chinese, the OpenCC-based processor converts traditional characters to simplified Chinese automatically.
- The skill generates two output files: a SubRip (.srt) subtitle file and a plain text (.txt) file.
- Progress is reported at 10% intervals to the console to ensure visibility during long processing tasks.
Video2txt: AI-Powered Media Transcription Setup
To use this within the Openclaw Skills framework, ensure you have Python 3.11 or 3.12 installed.
- Install the required Python dependencies:
python -m pip install -r requirements.txt
Ensure ffmpeg and ffprobe are installed and available in your system PATH.
Run the skill for the first time to automatically download the Whisper model:
python video_to_text.py --input "your_video_file.mp4"
Video2txt: AI-Powered Media Transcription Data Schema & Taxonomy
The skill generates structured output based on the input filename. Below is the metadata and file organization schema:
| File/Directory | Description | Format |
|---|---|---|
<filename>.srt |
Timestamped subtitle file | SubRip Subtitle |
<filename>.txt |
Full transcript text | Plain Text (UTF-8) |
models/ |
Storage for downloaded Whisper models | AI Model Binary |
Users can customize output locations using the --output-dir or --output-path flags to manage large datasets effectively.
Video2txt: AI-Powered Media Transcription Advanced Features
- GPU Acceleration: Support for CUDA devices to dramatically speed up transcription for long-form content.
- Model Customization: Choose from multiple Whisper model sizes (base, small, medium, large-v3) to balance accuracy and resource usage.
- VAD Filtering: Integrated Voice Activity Detection to filter out silence and non-speech noise.
- Background Execution: Optimized for Openclaw Skills background processing to prevent blocking the main interface.
- Fine-tuned Decoding: Adjustable beam size and compute types (e.g., int8, float16) for power users to optimize inference performance.
SKILL.md
Loading
Related Openclaw Skills
Chiikawa Persona Engine
A sophisticated multi-persona roleplay engine that replicates the unique verbal and non-verbal behaviors of the Chiikawa anime universe.
media-info
A local MediaInfo CLI integration for deep metadata extraction and orientation classification without root privileges.
Glass Cut & Shatter Video Generator
A specialized AI skill for creating high-quality vertical ASMR videos of glass cutting, cracking, and shattering.
Douyin Live AI Danmaku Assistant
A comprehensive real-time danmaku collection and AI-driven reply system for Douyin live streams powered by DeepSeek API.
agt0
agt0 is a unified local-first storage solution for AI agents that combines a SQLite database, a virtual filesystem, and long-term memory into a single .db file.
Build Opportunity Scout
An automated analysis skill that converts raw trend data into prioritized, actionable build opportunities for new applications and skills.
METADATA
Featured*
