A high-performance speech-to-text skill leveraging Whisper models for fast, multi-language audio transcription and translation via the command line.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install speech-to-text
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 speech-to-text 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 Speech-to-Text?
The Speech-to-Text skill for Openclaw Skills provides a robust interface for transcribing audio files using state-of-the-art Whisper models. By integrating with the inference.sh CLI, developers can execute complex transcription tasks, ranging from basic voice note conversion to large-scale meeting documentation with high accuracy and speed.
This skill is built for versatility, supporting over 99 languages and offering features like automatic translation to English and timestamped segments. Whether you are building an automated captioning tool or a searchable archive of voice recordings, this component of Openclaw Skills simplifies the path from raw audio to structured text data.
Speech-to-Text Use Cases
- Automated transcription for business meetings and video conferences to improve record-keeping.
- Generating SRT or VTT subtitle files for video content and social media clips.
- Transcribing podcast episodes to create SEO-friendly blog posts or searchable transcripts.
- Converting voice notes and interviews into text for qualitative research and data analysis.
- Enabling accessibility features by providing text alternatives for audio-only media.
How Speech-to-Text Works
- The user provides an audio source URL or localized file path through the CLI.
- The Openclaw Skills runner initiates a request to either the Fast Whisper V3 or Whisper V3 Large model via the inference.sh infrastructure.
- The audio is processed through a neural network that identifies the language and transcribes the speech content.
- If requested, the system performs an English translation or generates timestamped segments for every sentence.
- The final output is returned as a structured JSON object containing the full text and metadata.
Speech-to-Text Setup
Install the inference.sh CLI and authenticate to begin using this skill:
curl -fsSL https://cli.inference.sh | sh && infsh login
Once authenticated, you can add the skill to your project environment. To run a basic transcription immediately using the command line interface, use the following command:
infsh app run infsh/fast-whisper-large-v3 --input '{"audio_url": "https://example.com/audio.mp3"}'
Speech-to-Text Data Schema & Taxonomy
The skill outputs data in a JSON format designed for easy integration into other workflows within Openclaw Skills.
| Field | Type | Description |
|---|---|---|
text |
String | The full transcription of the audio content. |
segments |
Array | Optional list of objects containing start, end, and text for timestamps. |
language |
String | The ISO code of the detected or specified language. |
task |
String | Specifies whether the operation was transcribe or translate. |
Speech-to-Text Advanced Features
- Multi-model selection allowing users to prioritize between extreme speed (Fast Whisper) or maximum accuracy (Whisper V3 Large).
- Direct translation capabilities that convert foreign language audio directly into English text.
- Integration with video-audio-extractor to handle transcription directly from video files.
- Granular timestamp generation for precise subtitle alignment and navigation.
- Support for high-concurrency batch processing of large audio libraries via the CLI.
SKILL.md
Loading
Related Openclaw Skills
Social Media Carousel Agent Skill
A developer-centric skill for generating high-conversion multi-slide social media carousels using HTML-to-image automation and AI-driven design rules.
SEO Content Brief Skill
A data-driven automation skill for creating comprehensive SEO content briefs using real-time SERP analysis and keyword research.
Python Code Executor
A secure, sandboxed execution environment for Python scripts featuring over 100 pre-installed libraries for data science, web scraping, and media processing.
Prompt Engineering Guide
A masterclass in prompt engineering techniques for LLMs, image generators, and video models delivered through the Openclaw Skills ecosystem.
Storyboard Creation
An AI-driven workflow for generating cinematic storyboards and shot lists using CLI tools.
Talking Head Production
Automate the creation of high-quality talking head videos featuring AI avatars, synchronized lipsync, and professional voiceovers.
METADATA
Featured*
