A local, GPU-accelerated tool for converting audio files into SRT subtitles using OpenAI Whisper models without subscription fees.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install whisper-gpu-transcriber-skill
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 whisper-gpu-transcriber-skill 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 Whisper GPU Audio Transcriber?
Whisper GPU Transcribe is a powerful local processing tool designed to bridge the gap between high-quality transcription and data privacy. By leveraging OpenAI Whisper models, it provides a robust, offline solution for generating SRT subtitles from various audio formats. This skill is a core part of the Openclaw Skills library, enabling developers and creators to handle transcription tasks without relying on expensive, cloud-based SaaS platforms.
The utility is unique in its broad hardware support, automatically detecting and utilizing GPU acceleration across NVIDIA, AMD, Intel, and Apple Silicon. Whether you are working on high-end workstations or lightweight laptops, it optimizes the transcription process to deliver fast, accurate results while keeping your data entirely on your local machine.
Whisper GPU Audio Transcriber Use Cases
- Content creators seeking a free, high-quality alternative to paid captioning tools like CapCut.
- Professionals needing to convert meeting recordings into searchable text documents.
- Podcasters and educators looking to generate accurate subtitles for multi-language audio content.
- Developers integrating local speech-to-text workflows into larger automated pipelines via Openclaw Skills.
How Whisper GPU Audio Transcriber Works
- The AI agent triggers the transcription script when an audio file is detected or specified.
- The system performs an automatic hardware check to identify available GPU acceleration (CUDA, ROCm, Metal, or XPU).
- The specified Whisper model (defaulting to turbo for speed and accuracy) is loaded into memory.
- The audio stream is processed locally, extracting text and timestamp data.
- An SRT subtitle file is generated and saved directly to the source directory for immediate use.
Whisper GPU Audio Transcriber Setup
Ensure you have Python 3.8+ installed. Install the core dependencies using the following commands based on your hardware:
# For NVIDIA GPU users
pip install torch --index-url https://download.pytorch.org/whl/cu121nn# For Intel Arc GPU users
pip install torch==2.10.0+xpu
# Standard Whisper installation
pip install openai-whisper
Note: The first execution will automatically download the required model weights (approx. 1.5GB for the turbo model) to your local cache.
Whisper GPU Audio Transcriber Data Schema & Taxonomy
Whisper GPU Transcribe organizes its outputs and caches following a transparent local taxonomy:
| Item | Location | Description |
|---|---|---|
| Output File | same-as-source.srt | The generated subtitle file with precise timestamps. |
| Model Cache | ~/.cache/whisper | Storage for downloaded weights (tiny, base, turbo, etc.). |
| Execution | scripts/transcribe.py | The primary logic handler within the Openclaw Skills structure. |
Whisper GPU Audio Transcriber Advanced Features
- Multi-device acceleration support including Intel XPU and Apple Metal.
- Dynamic model switching allowing users to choose between 'tiny' for speed or 'large-v3' for maximum accuracy.
- Multi-language support with automatic language detection or manual override commands.
- Offline operation ensuring total privacy for sensitive meeting or personal recordings.
SKILL.md
Loading
Related Openclaw Skills
SEO Scout Pro
A specialized tool for discovering submission pages and analyzing form structures to streamline backlink building campaigns.
PDF Processing Skill
A comprehensive toolkit for programmatic PDF manipulation, data extraction, and document generation using industry-standard Python libraries and CLI utilities.
Generate Reading Activities
An automated workflow to generate comprehensive reading guides, study schedules, and assessment question banks based on any book title.
Psychology Assessment Creator
A professional framework for end-to-end management of psychological assessment content, from theoretical planning to data-driven performance optimization.
Daily Recommend
A sophisticated curation agent that delivers personalized daily recommendations of poetry, classic quotes, books, and films at optimal intervals.
PCS Hub API Integration
A professional toolset for designing and implementing PancakeSwap Hub swap functionality into external wallet apps, webviews, and headless bots.
Featured*
