DeepSeek Harness plugin

VoiceLens

Voice input and audio transcription for DeepSeek Harness and any skill-reading agent: a composer mic button (browser speech), a transcribe_audio tool, and a harness-agnostic `voicelens` CLI with

Jump to install

Source facts

Repository
EthanHuangEbor/VoiceLens
Latest update
Aug 14, 2026
Category
Workflow & Automation
GitHub stars
0
Format
plugin
Catalog evidence
Upstream dsh.bundle evidence
Evidence path
package.json#dsh.bundle
Checked against
0.1.0-rc.8
Upstream check date
2026-08-20

This evidence comes from the upstream catalog. This site has not installed, run, or security-reviewed the plugin.

Install

Start with a prompt that asks an agent to review the GitHub repository and source. Switch to the command if you want to install it yourself.

Copy this prompt into DSH, Codex, or another agent and ask it to review the GitHub repository and source first.

Do not install or run any commands yet. Read this plugin's GitHub repository, README, and relevant source code. Then answer the questions below clearly and directly so I can decide whether it fits my needs:

1. What is this plugin, and what problem does it solve?
2. Who is it for, and what are its typical use cases?
3. How is it used after installation? Include one minimal example.
4. What known limitations or privacy, security, compatibility, or maintenance risks does it have?
5. Give a clear recommendation: recommend, conditionally recommend, or do not recommend, with reasons.

Distinguish statements documented by the repository, inferences from source code, and unknowns. If evidence is insufficient, say so explicitly. Do not guess or simply repeat the README.

GitHub: https://github.com/EthanHuangEbor/VoiceLens
Plugin: VoiceLens
Author: EthanHuangEbor

Check the source files

Read the README and other files from this plugin directory before installing.

File explorer3 files
README.mdSource · read only

<p align="center"><img src="docs/assets/banner.png" alt="voicelens banner" width="100%"></p>

<h1 align="center">voicelens 🎙️</h1> <p align="center"><b>给文本模型装上「耳朵」——说人话,AI 就听得懂了。</b></p>

<p align="center"> <a href="https://github.com/EthanHuangEbor/VoiceLens/blob/main/LICENSE"><img src="https://img.shields.io/badge/license-MIT-green" alt="License"></a> <img src="https://img.shields.io/badge/node-%3E%3D22-blue" alt="Node"> <img src="https://img.shields.io/badge/DSH%20plugin-native-8b5cf6" alt="DSH native plugin"> </p>

---

为什么做这个

跟 AI 聊天,绝大多数时候你还是得打字。

想躺着聊?想一边喝咖啡一边口述一段思路?想偷懒?—— 打开输入框,旁边多了个小喇叭,点一下,直接说。每停顿一下,刚说的那句话就自动落进输入框,说到哪儿你都看得见。

就这么简单,不用训练,不用改习惯。

它现在能干嘛

  • 🎙️ 边说边出字:说话时字词实时浮现在喇叭上方,每断一句立刻写入输入框——不再是"录完了才一次性蹦出来一大段"。
  • 📝 模型也能转写音频:扔给它一个音频文件或链接,transcribe_audio 工具直接转成文字证据。
  • 🖥️ 不只在 DSH 能用:同一个引擎打包成了 voicelens 命令行,Claude Code / Codex / Pi / OpenClaw 里的 agent 读同一份技能就能用。
  • 🔌 转写引擎随便换:Groq(免费)/ OpenAI 兼容端点 / 本地 whisper.cpp,一份配置,所有宿主共享。

<p align="center"><img src="docs/assets/demo.png" alt="DSH 输入框 + voicelens 麦克风按钮" width="720"></p>

快速上手

# DSH 原生插件(麦克风按钮 + transcribe_audio 工具)
dsh plugin --profile web add voicelens

# 或装 CLI(任何宿主都能用)
npm install -g voicelens

装完刷新页面,输入框左边就是那个小喇叭。点一下,说话。

浏览器支持(实话实说)

浏览器体验
Chrome / Edge✅ 零配置,浏览器原生语音识别,开箱即用
Firefox / Safari✅ 也能用,走「录音 → Host 转写」通道,需要先配一个转写引擎(见下)
其他只要支持麦克风录音,基本都能跑

配置转写引擎(给 Firefox 和文件转写用)

export VOICELENS_PROVIDER=groq
export VOICELENS_GROQ_API_KEY=gsk_...   # 免费申请,个人用绰绰有余

Chrome / Edge 用户这一步可以跳过——浏览器自带识别,不要钱不要 key。

架构:站在巨人的肩膀上

一句话:把 modlens 那套「一个引擎、多端适配」的思路,从视觉搬到了语音。

<p align="center"><img src="docs/assets/architecture.png" alt="voicelens architecture" width="720"></p>

modlens 证明了一件事:核心引擎(CLI + 多 provider)+ 一份通用技能 + 各平台薄薄一层原生适配,就能让一个能力到处跑。voicelens 照搬了这个套路,只是把「读图」换成了「听话」。

唯一绕不开的差别是:图片能靠「粘贴」进入所有宿主,语音没有这条现成通道——所以 DSH 里多了个麦克风按钮,这是必须自己补的一小块。模型侧的文件转写,则是天然的多平台。

致谢

这个项目能长出来,得真心谢谢两个项目:

  • modlens —— 架构的蓝本。三层分离、provider 接口、分层配置、doctor 诊断,几乎都是跟它学的。作者 @liustack 的思路,我"抄"得很开心,也抄得理直气壮——好设计就该被复用。
  • DeepSeek Harness (DSH) —— 开放的插件体系(bundle/client 清单、Slots 插槽、llm 服务……)让「语音输入」能做成一个干净的原生插件,而不是一堆 hack。

没有这俩项目,就没有 voicelens。Respect。🙏

License

MIT,随便用,别客气。