← Back to Skills Marketplace
396
Downloads
0
Stars
0
Active Installs
2
Versions
Install in OpenClaw
/install u2-audio-file-transcriber
Description
Transcribe audio files via UniCloud ASR (云知声语音识别, recorded audio → text) API from UniSound. Supports multiple formats, optimized for finance, customer servic...
Usage Guidance
This package appears to be a straightforward UniSound ASR client. Before installing: (1) only provide UNISOUND_APPKEY/UNISOUND_SECRET from a trusted UniSound account and avoid using production-sensitive audio with the provided UAT test credentials; (2) verify or set UNISOUND_BASE_URL if you intend to use a production endpoint instead of the included UAT host; (3) understand that the skill uploads audio to UniSound servers (no local transcription or fallback is allowed by the SKILL.md); (4) review the script if you need stricter privacy controls (e.g., change endpoints, remove hard-coded vocab_id, or add encryption in transit if required). If any of these are unacceptable (uploading sensitive audio to an external ASR), do not use the skill.
Capability Analysis
Type: OpenClaw Skill
Name: u2-audio-file-transcriber
Version: 1.0.1
The skill facilitates audio transcription via the UniSound API but contains significant security vulnerabilities. It uses an unencrypted HTTP endpoint (http://af-asr.uat.hivoice.cn) for uploading audio files, which risks exposing sensitive data to interception. Additionally, SKILL.md provides hardcoded UAT test credentials (AppKey/Secret), and scripts/transcribe.py contains a hardcoded vocab_id. While the behavior aligns with the stated purpose, the lack of transport encryption and inclusion of credentials warrant a suspicious classification.
Capability Assessment
Purpose & Capability
Name and description (UniSound ASR audio transcriber) match the included script and the declared requirements (python3 and UniSound API credentials). The script communicates only with UniSound endpoints (configurable base_url) and implements an upload→transcribe→poll workflow appropriate for the stated purpose.
Instruction Scope
SKILL.md restricts the agent to calling the provided script and to never attempt local transcription or alternate services; this is coherent with a policy that forces use of the UniSound API, but it is prescriptive. The SKILL.md and script both upload audio to UniSound servers (the doc states no third-party endpoints are used). The instructions intentionally prevent fallback/local processing — not a security flaw but an operational restriction the user should understand.
Install Mechanism
There is no installer in the registry metadata (instruction-only). The README directs users to pip install the small requirements.txt (requests and urllib3) — a standard, low-risk dependency installation from well-known Python packages.
Credentials
Declared required env vars are UNISOUND_APPKEY and UNISOUND_SECRET (primaryEnv = UNISOUND_SECRET), which are appropriate for an API client. The runtime code may also read optional env vars such as UNISOUND_BASE_URL, UNISOUND_USERID, UNISOUND_AUDIOTYPE, and UNISOUND_USE_HOT_DATA; these are reasonable but not listed in requires.env. The number and sensitivity of required variables are proportionate to the task.
Persistence & Privilege
The skill does not request persistent/always-on inclusion (always:false) and does not modify other skills or system-wide settings. It simply runs a CLI script when invoked.
How to Use
- Make sure OpenClaw is installed (local or Docker)
- Run the install command in chat:
/install u2-audio-file-transcriber - After installation, invoke the skill by name or use
/u2-audio-file-transcriber - Provide required inputs per the skill's parameter spec and get structured output
Version History
v1.0.1
# v1.0.1 Changelog
- Removed all documentation and reference files: README.md and all content in references/.
- SKILL.md is now the primary and only documentation, with essential usage and configuration info.
- No user-facing changes to runtime usage or API behavior; update is documentation-only.
- Reduced redundant or sample guidance—relevant instructions are now consolidated.
v1.0.0
u2-audio-file-transcriber v1.0.0
- Initial release: transcribe recorded audio files to text using the UniCloud ASR (云知声语音识别) API.
- Supports multiple audio formats (WAV, MP3, M4A, FLAC, OGG), up to 2 hours or 100MB per file.
- Optimized domains include finance and customer service.
- Script-based CLI usage with options for output format, audio type, and domain selection.
- Requires configuration of API credentials via environment variables; includes UAT test credentials for evaluation.
- Designed only for pre-recorded audio file transcription, not real-time, streaming, or TTS use cases.
Metadata
Frequently Asked Questions
What is U2-audio-file-transcriber?
Transcribe audio files via UniCloud ASR (云知声语音识别, recorded audio → text) API from UniSound. Supports multiple formats, optimized for finance, customer servic... It is an AI Agent Skill for Claude Code / OpenClaw, with 396 downloads so far.
How do I install U2-audio-file-transcriber?
Run "/install u2-audio-file-transcriber" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.
Is U2-audio-file-transcriber free?
Yes, U2-audio-file-transcriber is completely free, licensed under MIT-0. You can download, install and use it at no cost.
Which platforms does U2-audio-file-transcriber support?
U2-audio-file-transcriber is cross-platform and runs anywhere OpenClaw / Claude Code is available (cross-platform).
Who created U2-audio-file-transcriber?
It is built and maintained by aaiccee (@aaiccee); the current version is v1.0.1.
More Skills