🎤 Transcribe audio files using Qwen ASR. 千问STT
ReviewAudited by ClawScan on May 10, 2026.
Overview
This appears to be a straightforward transcription tool, but it uploads the chosen audio to a Qwen demo service and may print the uploaded-audio link.
This skill is reasonable for ordinary transcription tasks. Before installing, understand that it sends the selected audio to a third-party Qwen demo endpoint and may print the uploaded-audio URL in logs. Avoid using it for recordings containing highly sensitive personal, business, or confidential information unless you are comfortable with that external processing.
Findings (3)
Artifact-based informational review of SKILL.md, metadata, install specs, static scan signals, and capability signals. ClawScan does not execute the skill or run runtime probes.
Private or sensitive recordings may be shared with the external demo service during transcription.
The selected audio bytes are uploaded to a remote Qwen demo service for transcription. This is purpose-aligned, but it means the audio leaves the user's environment.
BASE_URL = "https://qwen-qwen3-asr-demo.ms.show" ... form.add_field("files", file, filename="audio") ... res = await api_request("/gradio_api/upload", data=form)Use this only for audio you are comfortable sending to that service; use a local ASR tool for highly sensitive recordings.
Anyone who can see the logs or tool output may see a link associated with the uploaded audio.
After upload, the script logs the remote audio URL, which may appear in local logs or tool output.
_LOGGER.warning("Audio file: %s", audio_url)Avoid sharing logs from transcription runs, and consider redacting or removing this logging if maintaining the skill.
Future package changes or package-index issues could affect what code is installed to run the skill.
The uv script declares Python dependencies without pinned versions, so runtime/install behavior depends on external package resolution.
# dependencies = ["aiohttp", "argparse", "gradio_client"]
Prefer pinned versions or a lockfile for reproducible installs, especially in sensitive environments.