A high-performance audio transcription tool that converts Chinese and English speech into text using the StepFun streaming API.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install step-asr
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 step-asr 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 Step ASR?
Step ASR is a versatile speech-to-text utility designed to process audio files through the StepFun ASR streaming API. It excels in delivering low-latency transcriptions for both Mandarin and English, making it an essential addition to any Openclaw Skills collection. By leveraging HTTP SSE streaming, it provides real-time text feedback, allowing developers to monitor transcription progress as it happens.
This skill supports a wide range of audio formats, including PCM, WAV, MP3, and OGG/OPUS. It also includes advanced capabilities like terminology correction via prompts, ensuring high accuracy for domain-specific vocabulary. Whether you are building automated transcription pipelines or need a quick CLI tool for voice data, Step ASR provides the reliability and speed required for modern AI workflows.
Step ASR Use Cases
- Real-time transcription of meeting recordings or voice memos.
- Automated subtitle generation for video content in Chinese or English.
- Improving transcription accuracy for technical jargon using custom prompts.
- Batch processing audio files for data analysis or archival purposes.
- Integrating speech-to-text capabilities into larger Openclaw Skills automation scripts.
How Step ASR Works
- The user provides an audio file path and optional parameters like language or prompt.
- The skill identifies the audio format and sample rate, either automatically or via user flags.
- A connection is established with the StepFun API using HTTP Server-Sent Events (SSE).
- Audio data is streamed to the server, and transcription results are received in real-time.
- The final text is output to the console or saved to a specified file, optionally in JSON format with usage statistics.
Step ASR Setup
To use this skill, ensure you have Python 3 installed and a valid API key from StepFun.
First, set your environment variable:
export STEPFUN_API_KEY="your_api_key_here"
Alternatively, add it to your configuration for Openclaw Skills:
{
"skills": {
"step-asr": {
"apiKey": "YOUR_STEPFUN_API_KEY"
}
}
}
Run the transcription script using the following command:
python3 scripts/transcribe.py /path/to/audio.wav
Step ASR Data Schema & Taxonomy
The skill handles various audio inputs and can generate text or structured JSON outputs for seamless integration into other Openclaw Skills.
| Data Type | Description |
|---|---|
| Input Formats | .wav, .mp3, .pcm, .raw, .ogg, .opus |
| Output Formats | Plain text (.txt) or structured JSON |
| Metadata | Includes model name, language, and token usage statistics in JSON mode |
Step ASR Advanced Features
- Terminology correction hints via the --prompt flag to handle industry-specific terms.
- Second-pass error correction for improved linguistic accuracy.
- Inverse text normalization (ITN) to format numbers, dates, and currency correctly.
- Flexible output options including raw console streaming or structured JSON files for Openclaw Skills integration.
- Support for raw PCM files with customizable sample rates and codec settings.
SKILL.md
Loading
Related Openclaw Skills
Playwright MCP
A high-performance browser automation skill that enables AI agents to interact with websites, fill forms, and extract data via the Playwright framework.
Self-Evolving Agent
A meta-analytical automation tool that observes AI agent performance and suggests optimized instruction updates based on semantic log patterns and real-time monitoring.
OpenClaw Self-Healing System
An autonomous 4-tier recovery system that monitors gateway health and uses AI to diagnose and repair persistent failures.
MemoryBox
MemoryBox is a zero-dependency CLI tool designed to maintain a lean 3-tier hierarchical memory structure for Openclaw Skills.
Skill Deps Doctor
A cross-platform dependency validator designed to ensure all system-level and project-level requirements are met for Openclaw Skills.
TA Analyzer
A professional-grade technical analysis engine that calculates indicators, detects chart patterns, and generates trading signals using CCXT market data.
Featured*
