🎤 Transcribe audio files using Qwen ASR. 千问STT

Transcribe audio files using Qwen ASR (千问STT). Use when the user sends voice messages and wants them converted to text.

Security Scan

The skill's code and instructions match its stated purpose (upload audio to the Qwen ASR demo and return a transcript); it requests no credentials, but it will upload user audio to a third-party demo service (privacy consideration).

Purpose & Capability

Name, description, SKILL.md and scripts/main.py all describe transcribing audio via the Qwen ASR demo (BASE_URL = https://qwen-qwen3-asr-demo.ms.show). The script's network calls and use of gradio_client are consistent with that purpose. Requiring the 'uv' runner matches the provided usage examples (uv run ...).

Instruction Scope

Runtime instructions tell the agent to run scripts/main.py with uv and/or pipe audio over stdin. The script uploads the audio to the Qwen ASR demo (/gradio_api/upload) and then calls the service via gradio_client to get transcripts. This is within the stated purpose, but it does mean user audio is transmitted to an external third-party endpoint — the SKILL.md mentions the demo host but does not explicitly warn about uploading potentially sensitive audio.

Install Mechanism

Install entries reference installing 'uv' (brew) and pip alternatives for uv/aiohttp/gradio_client. These are reasonable and expected for running the Python script. Minor inconsistency: registry metadata lists a brew requirement for 'uv', while SKILL.md metadata also includes pip-based install options; both are plausible but the mismatch is worth noticing.

Credentials

The skill declares no required environment variables, credentials, or config paths and the code does not read secrets or unrelated environment variables. No credentials are requested or needed for the demo service, which is proportional to the functionality.

Persistence & Privilege

always is false and the skill does not modify other skills or system-wide configuration. It does not request permanent presence or elevated privileges.

Assessment

This skill appears to do what it says: it uploads audio to the Qwen ASR demo (https://qwen-qwen3-asr-demo.ms.show) and returns the transcript. Before installing, consider privacy and compliance: any audio you transcribe will be sent to that third-party demo service (no API key required). If the audio is sensitive, do not use this skill or run it only against a trusted/private ASR endpoint. Also note the skill expects the 'uv' runner and Python dependencies (aiohttp, gradio_client); you may prefer to run the included script locally in an isolated environment to inspect behavior and avoid installing system-wide packages.

Qwen ASR

Transcribe an audio file (wav/mp3/ogg...) to text using Qwen ASR. No configuration or API key required.

Usage

uv run scripts/main.py -f audio.wav
cat audio.wav | uv run scripts/main.py > transcript.txt

About

Qwen ASR is a free and open-source speech-to-text model. It is trained on a large dataset of audio files from the web. It is available in multiple languages. This skill bases on the Qwen ASR Demo service (qwen-qwen3-asr-demo.ms.show).