3274375092/dsh-voice 预览 preview

3274375092/dsh-voice

DeepSeek Harness 的语音输入插件:麦克风 → 本地/浏览器语音识别 → 文本作为普通聊天消息提交。仅输入且与预设无关。

Project Overview项目介绍

To install the base plugin, run dsh plugin --profile web add @nn12138/dsh-voice in your terminal. dsh-voice is a native voice input plugin built exclusively for DeepSeek Harness. After installation, it adds a microphone button to the left of the web UI’s composer bar, and registers a configurable global hotkey (default Ctrl+Space) that triggers voice capture. It does not alter system prompts or agent presets, so it works as an additional input method across all DSH preset types including code, standard, and custom presets.

When you need to input spoken content instead of typing, just click the button or press the hotkey to start recording. It streams partial recognition results to the UI while you speak, and automatically commits the final full transcript after you stop speaking, detected via voice activity detection (VAD). It works out of the box with zero configuration by default, using an adaptive dual-engine setup that falls back to the browser’s built-in Web Speech API if no local offline model is available. This makes it ideal for any DSH user who wants a quick way to input spoken text without manual typing.

If you want offline, private local recognition, you need to install the optional sherpa-onnx-node runtime dependency, then run the dsh-voice-models command to download the ~100MB ASR model files to a local directory. You can customize the hotkey, VAD threshold, tail padding duration, and engine preference via the DSH profile configuration file, and you will need to restart the dsh web service after changing any settings for changes to take effect. The project is open source under the permissive MIT license, and all installation steps avoid running risky lifecycle scripts for security.

这是一款专为 DeepSeek Harness 开发的原生语音输入插件。它会在 DSH 的网页界面左侧添加麦克风按钮,同时支持可配置的全局热键(默认为 Ctrl+空格),让用户可以通过语音直接生成聊天文本,无需手动输入。插件不会修改系统提示词或代理预设,适配所有类型的 DSH 对话与代码预设,相当于额外增加了一种语音输入方式。

用户在对话过程中需要口述内容、不便手动打字时,点击界面上的麦克风按钮或按下预设热键即可开始录音,说话过程中会实时流式显示识别结果,结束说话后会自动提交完整文本到对话输入框。它默认开箱即用,采用自适应双引擎方案,无本地模型时自动回退到浏览器内置语音识别,适合需要快速输入口述内容的 DSH 用户日常使用。

如果需要离线隐私的本地识别,用户可以额外安装 sherpa-onnx-node 运行时并运行 dsh-voice-models 命令下载约 100MB 的模型文件,之后可通过配置文件修改热键、VAD阈值等参数。修改配置后需要重启 dsh web 服务生效,项目采用 MIT 许可证开源,可免费使用和修改。

Pre-install check安装前体检Compatibility · Security兼容性 · 安全性 1 warning1 项注意
  • Only 8 stars - very few users, little community feedback星标只有 8,几乎没人在用,遇到问题缺少社区反馈
DSH walks through these 9 checksDSH 会逐条核对这 9 项

Compatibility兼容性

  • DSH, Node, OS and profile requirementsDSH 版本 / Node 版本 / 操作系统 / profile 是否满足要求
  • External dependencies and runtimes (Electron / Python / Docker, ...)外部依赖与运行时(Electron / Python / Docker 等)是否齐备
  • Conflicts with installed plugins: command names, skill / tool names, ports, duplicate MCP registration与已装插件是否冲突:命令名、skill / tool 重名、端口占用、重复 MCP 注册

Security安全性

  • Repo matches the facts registered here; archived or abandoned?仓库是否与页面登记一致,是否归档或长期停更
  • Safety of preinstall / install / postinstall and install.sh / setup.ps1preinstall / install / postinstall 与 install.sh、setup.ps1 是否安全
  • curl|bash, download-then-execute, obfuscation, unrelated domains → stop immediatelycurl|bash、下载即执行、混淆代码、无关域名 → 立刻停止
  • Typosquatting or unmaintained packages among the new dependencies新增依赖里有没有 typosquatting 或无人维护的包
  • Requested permissions vs. what the feature actually needs申请了哪些权限、是否超出功能所需(filesystem / network / shell / clipboard)
  • Any sudo / admin requirement, plus uninstall and rollback是否要求 sudo / 管理员权限,以及卸载与回滚方式

Anything uncertain must be marked unknown with a note on how to confirm it. This site's signal screen is a static snapshot, not a security audit.拿不准的必须标「未知」并说明要我怎么确认。本站的信号筛查是静态快照,不能替代安全审计。

Or use CLI install (for developers)或使用命令行安装(适合开发者)

CLI Install命令行安装

dsh plugin --profile web add @nn12138/dsh-voice

把 3274375092/dsh-voice 加入你的 DSH 配置(web profile)即可启用。

READMEREADME

dsh-voice 🎤

English | 中文

A voice input plugin for DeepSeek Harness: click 🎤 in the web UI (or press a hotkey), speak, and the recognized text is submitted as a normal chat message. Input only — it never touches the agent preset/persona, so it behaves like "another input method" in every mode.

npm License

Features

  • 🎤 Voice input: microphone button (platform design-system UI) + configurable global hotkey (Ctrl+Space by default)
  • ⚡ Live recognition: streaming partial transcripts are echoed while you speak; VAD finalization commits on stop (0.6s tail padding keeps sentence endings)
  • 🧠 Adaptive dual engine: host-native ASR (sherpa-onnx-node zipformer2 + silero VAD, offline/private) with automatic fallback to browser Web Speech (zero extra dependencies)
  • 🔌 Preset-agnostic: does not touch the persona/system prompt; works with code/standard/minimal/custom presets
  • 📦 Optional models: zero-config out of the box; run dsh-voice-models when you want native offline recognition

Installation

# Plugin
dsh plugin --profile web add @nn12138/dsh-voice

# Optional: offline native recognition (the plugin does not auto-install this runtime)
dsh plugin --profile web add sherpa-onnx-node
dsh-voice-models            # one-shot model download (~100MB) → ./dsh-voice-models

# Optional: configuration (edit ~/.dsh/profiles/web/cordis.patch.yml)
- id: voice
  config:
    modelDir: './dsh-voice-models'   # native ASR model directory
    hotkey: 'ctrl+space'             # global hotkey
    vadThreshold: 0.3                # lower = less clipping at sentence boundaries
    tailPadSeconds: 0.6              # tail-padding duration
    engine: auto                     # auto (default) | native | browser

The row-level config is received by the host half. engine and hotkey are synced to the browser half over the /voice.config loopback RPC, so there is no separate client config to write. auto probes host native capability: with a model it uses native; without one it falls back to Web Speech, so zero-config users keep working. Restart dsh web after changing the config.

Showing the opening section of the README — the full document lives in the repository以上为 README 开头摘要,完整文档在仓库内 · View the full README on GitHub →在 GitHub 查看完整 README →

← 上一个 Prev dsh-director-toolkit 下一个 Next dsh-research →