← 返回 Skills 市场
Lx Whisper Transcribe
作者
liuxuebin20260309
· GitHub ↗
· v1.1.0
· MIT-0
41
总下载
0
收藏
0
当前安装
1
版本数
在 OpenClaw 中安装
/install lx-whisper-transcribe
功能描述
OpenClaw语音转文字技能 - 使用whisper.cpp实现本地语音识别,完全本地处理,无需API密钥,支持中文。
使用说明 (SKILL.md)
lx-whisper-transcribe
OpenClaw语音转文字技能 - 使用whisper.cpp实现本地语音识别
描述
此技能解决OpenClaw升级后语音消息无法被转录为文字的问题。通过使用whisper.cpp和国内镜像源,提供完全本地化的中文语音转文字功能,无需API密钥。
功能
- ✅ 接收并转录语音消息为文字
- ✅ 支持中文语音识别
- ✅ 使用本地模型,无需API密钥
- ✅ 在网络受限环境下通过国内镜像源正常工作
触发条件
当用户提到“语音转文字”、“转录音频”、“语音识别”或类似关键词时,应激活此技能。
使用方法
whisper /path/to/audio.ogg
注意事项
- 首次使用需要下载模型文件
- 需要预先编译 whisper.cpp
- 需要安装 ffmpeg
安全使用建议
Before installing, review the wrapper commands and only proceed if you are comfortable creating ~/bin/whisper, changing PATH for future shell sessions, and downloading models from hf-mirror.com or Hugging Face. Consider using a temporary PATH change first, verify which command runs with command -v whisper, and remove the ~/.bashrc line if you no longer want the wrapper active.
能力评估
Purpose & Capability
The requested capabilities fit a Whisper transcription skill: reading audio files, running ffmpeg/wget/curl, and caching models are purpose-aligned. The documentation is somewhat inconsistent because it calls the setup fully local/offline while also describing first-use model downloads from mirror sites.
Instruction Scope
The declared triggers are mostly specific, but SKILL.md also says to activate on similar keywords, which could invoke transcription more broadly than intended.
Install Mechanism
Installation is manual and visible in SHARE_GUIDE.md, but it creates a ~/bin/whisper wrapper and appends ~/bin to PATH in ~/.bashrc without a strong upfront warning or removal steps.
Credentials
The permissions and external tools are proportionate for audio transcription and model retrieval; there is no evidence of credential access, broad local indexing, exfiltration, or destructive behavior.
Persistence & Privilege
The persistent PATH change and user-local wrapper can shadow other whisper commands in future shells, but this is disclosed in setup commands and is aligned with making the CLI available.
如何使用
- 确保已安装 OpenClaw(本地或 Docker 部署)
- 在对话框中输入安装命令:
/install lx-whisper-transcribe - 安装完成后,直接呼叫该 Skill 的名称或使用
/lx-whisper-transcribe触发 - 根据 Skill 的参数说明提供必要输入,即可获得结构化输出
版本历史
v1.1.0
- Adds support for OpenClaw offline Chinese speech-to-text using whisper.cpp.
- Fully local processing; no API key required.
- Utilizes domestic mirrors for downloads, improving usability in restricted network environments.
- Recognizes and transcribes audio with Chinese language support.
- Triggered by keywords such as “语音转文字”, “转录音频”, and “语音识别”.
元数据
常见问题
Lx Whisper Transcribe 是什么?
OpenClaw语音转文字技能 - 使用whisper.cpp实现本地语音识别,完全本地处理,无需API密钥,支持中文。 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件,目前累计下载 41 次。
如何安装 Lx Whisper Transcribe?
在 OpenClaw 或 Claude Code 对话框中运行命令「/install lx-whisper-transcribe」即可一键安装,无需额外配置。
Lx Whisper Transcribe 是免费的吗?
是的,Lx Whisper Transcribe 完全免费,采用 MIT-0 许可证,可自由下载、安装和使用。
Lx Whisper Transcribe 支持哪些平台?
Lx Whisper Transcribe 跨平台运行,可在任意部署了 OpenClaw / Claude Code 的环境中使用(cross-platform)。
谁开发了 Lx Whisper Transcribe?
由 liuxuebin20260309(@liuxuebin20260309)开发并维护,当前版本 v1.1.0。
推荐 Skills