← Back to Skills Marketplace
Fun-ASR-File
by
chenggongdu
· GitHub ↗
· v1.1.0
· MIT-0
123
Downloads
0
Stars
0
Active Installs
1
Versions
Install in OpenClaw
/install fun-asr-file
Description
阿里云百炼 FunASR 本地音频文件识别（非流式），使用阿里云 DashScope API 进行语音转文字。针对本地音频文件优化，支持自动格式转换，适合批量文件转写场景。
README (SKILL.md)
Fun-ASR-File
Fun-ASR 是通义实验室百聆团队推出的端到端语音识别大模型，是基于数千万小时真实语音数据训练而成，具备强大的上下文理解能力与行业适应性。
激活条件
| 触发场景 | 说明 |
|---|---|
| 用户发送音频文件 | .wav / .mp3 / .m4a / .flac / .ogg 等格式 |
| 用户要求转录 | "转写音频"、"语音转文字" |
| 音频文件处理 | 需要提取音频中的文字内容 |
配置
设置环境变量：
export DASHSCOPE_API_KEY="sk-xxx"
使用方法
安装依赖
pip install dashscope
# 如需自动格式转换，请安装 FFmpeg
转写音频文件
python scripts/cli.py audio.wav
# 支持 .wav / .mp3 / .m4a / .flac / .ogg 等格式
# 会自动转换为 API 要求的格式（16kHz, 单声道, pcm_s16le）
版本：1.1.0 创建于：2026-03-16 更新：2026-04-09 - 改为非流式调用，优化本地文件处理
Usage Guidance
Before installing or using this skill:
- Expect to provide an Aliyun DashScope API key (DASHSCOPE_API_KEY). The registry metadata omitted this; verify the publisher and update metadata before trusting the skill.
- Treat the API key as a secret. Use a least-privilege key and avoid pasting long-lived account keys unless you control the account and understand DashScope billing/auth.
- The skill requires the dashscope Python package and FFmpeg. Install packages from trusted sources and verify dashscope's provenance (PyPI project page, source repo) before pip installing.
- NOTES.md mentions integration with a 'tikhub-douyin-media-links' skill that would download media — if you plan to chain skills, be aware that downloaded content and the chain may transmit data to external services.
- If you need higher assurance, request that the publisher correct the registry metadata to declare DASHSCOPE_API_KEY as a required credential and provide a homepage or source repo for review; run the skill in a sandboxed environment first.
Capability Analysis
Type: OpenClaw Skill
Name: fun-asr-file
Version: 1.1.0
The skill is a legitimate implementation for audio-to-text transcription using Alibaba Cloud's DashScope API. The core logic in `scripts/cli.py` uses the official `dashscope` SDK to process local files and correctly handles API responses without any evidence of data exfiltration, unauthorized network calls, or malicious command execution.
Capability Assessment
Purpose & Capability
The name/description (local ASR via FunASR / DashScope) match the code and SKILL.md: the script calls the DashScope ASR API to transcribe local files and recommends FFmpeg for preprocessing. Functionality requested (local file transcription) is coherent with the implementation.
Instruction Scope
SKILL.md and scripts/cli.py limit actions to reading a user-supplied audio file, preprocessing instructions (FFmpeg), and sending the file to DashScope; they do not attempt to read unrelated system files. NOTES.md documents optional interaction with another skill (tikhub-douyin-media-links) for fetching media links — that implies possible cross-skill network download workflows if combined, but the skill itself does not autonomously perform such downloads.
Install Mechanism
This is instruction-only (no install spec). It requires the dashscope Python SDK and recommends FFmpeg; both are typical for this task. No suspicious external download URLs or archive extraction are present. The lack of an install spec means dependencies would be installed manually by the operator, which reduces automatic risk but requires the user to trust the dashscope package source.
Credentials
The SKILL.md and scripts/cli.py require DASHSCOPE_API_KEY (dashscope.api_key = os.environ.get('DASHSCOPE_API_KEY')), but the skill metadata declared no required environment variables or primary credential. That inconsistency is concerning: the runtime needs an API key but the registry entry does not advertise it. No other unrelated credentials are requested.
Persistence & Privilege
always:false and the code does not modify agent/system configuration or other skills. The skill runs only when invoked and does not request persistent elevated privileges.
How to Use
- Make sure OpenClaw is installed (local or Docker)
- Run the install command in chat:
/install fun-asr-file - After installation, invoke the skill by name or use
/fun-asr-file - Provide required inputs per the skill's parameter spec and get structured output
Version History
v1.1.0
改为非流式调用，优化本地音频文件处理，支持自动格式转换
Metadata
Frequently Asked Questions
What is Fun-ASR-File?
阿里云百炼 FunASR 本地音频文件识别（非流式），使用阿里云 DashScope API 进行语音转文字。针对本地音频文件优化，支持自动格式转换，适合批量文件转写场景。 It is an AI Agent Skill for Claude Code / OpenClaw, with 123 downloads so far.
How do I install Fun-ASR-File?
Run "/install fun-asr-file" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.
Is Fun-ASR-File free?
Yes, Fun-ASR-File is completely free, licensed under MIT-0. You can download, install and use it at no cost.
Which platforms does Fun-ASR-File support?
Fun-ASR-File is cross-platform and runs anywhere OpenClaw / Claude Code is available (cross-platform).
Who created Fun-ASR-File?
It is built and maintained by chenggongdu (@chenggongdu); the current version is v1.1.0.
More Skills