A high-performance speech recognition CLI that enables seamless audio transcription for AI agents within the Openclaw Skills library.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install asr-claw
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 asr-claw 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 asr-claw?
asr-claw is a robust speech-to-text utility designed to integrate deeply with the Openclaw Skills ecosystem. It provides AI agents with the specialized ability to process audio from various sources, including local files, web URLs, and real-time piped streams.
By unifying several state-of-the-art ASR engines under a single command-line interface, this tool allows developers to choose between local, privacy-focused models like Qwen and high-speed cloud services like Deepgram or OpenAI. It is optimized for automation, making it easy to build voice-enabled agents and transcription pipelines.
asr-claw Use Cases
- Transcribing live audio streams from Android devices when paired with adb-claw.
- Generating professional SRT or VTT subtitle files for video and lecture content automatically.
- Converting voice recordings and meetings into searchable, structured text for AI analysis.
- Monitoring live streams or audio feeds to trigger agent actions based on spoken keywords.
- Building custom voice-to-command interfaces for Openclaw Skills automation scripts.
How asr-claw Works
- The user selects and installs an ASR engine, such as the local qwen-asr or a cloud provider like OpenAI.
- Audio data is fed into the CLI via file path, URL, or standard input (stdin).
- The tool utilizes Voice Activity Detection (VAD) or fixed-time chunking to process the audio efficiently.
- The chosen engine transcribes the audio segments into text.
- asr-claw returns a structured JSON envelope containing full text, timestamps, and segment data for downstream processing.
asr-claw Setup
To begin using this tool within the Openclaw Skills framework, follow these steps:
# Install the default local engine (qwen-asr)
asr-claw engines install qwen-asr
# Configure your preferred language
claw config set asr-claw.default_lang en
# (Optional) Set up cloud engines by exporting API keys
export OPENAI_API_KEY=your_key_here
You can verify the installation by running asr-claw doctor to check platform dependencies and engine status.
asr-claw Data Schema & Taxonomy
The skill produces a standard JSON output envelope for all transcription tasks, ensuring compatibility across different Openclaw Skills.
| Key | Type | Description |
|---|---|---|
ok |
boolean | Indicates if the transcription was successful |
data.full_text |
string | The complete transcribed text |
data.segments |
array | List of objects containing index, start/end times, and segment text |
data.engine |
string | The ASR engine used for the task |
duration_ms |
integer | Total processing time in milliseconds |
Local configuration and cached models are stored in ~/.asr-claw/.
asr-claw Advanced Features
- Real-time streaming support with native VAD for low-latency transcription.
- Seamless integration with adb-claw for direct Android system audio capture.
- Multi-engine support including Local (Qwen, Whisper) and Cloud (Deepgram, Doubao, OpenAI).
- Flexible output formatting for text, JSON, and standard subtitle formats (SRT/VTT).
- Support for HuggingFace mirrors to facilitate model downloads in restricted network environments.
SKILL.md
Loading
Related Openclaw Skills
AutoHotkey Reference Skill
A local, high-performance reference tool for AutoHotkey development providing instant access to patterns, best practices, and migration guides.
Data Analyst Assistant
An automated data science companion for cleaning, analyzing, and visualizing datasets using Python and Pandas.
Browser Fingerprinting Reference
A comprehensive, offline reference tool for developers to quickly lookup browser fingerprinting concepts, security patterns, and debugging workflows.
Learning Check-in
A lightweight habit-tracking tool that helps users maintain daily learning consistency through simple check-ins and intelligent reminders.
Registry Sync
An automation utility that synchronizes local AI agent skills, workflows, and templates with a centralized Feishu registry spreadsheet.
BCC Reference Tool
A zero-dependency reference utility providing instant, offline access to BCC documentation and BPF coding patterns.
Featured*
