dsh-sensevoice-npu：在 Intel NPU 上跑 SenseVoiceSmall 的 DSH 语音转文字 Provider
dsh-sensevoice-npu 为 DeepSeek Harness 官方实验语音包新增 OpenVINO NPU 识别 provider，在不变更语音按钮与 speech API 的前提下运行 SenseVoiceSmall。插件不下载模型，需 Python 3.11–3.13 与 NPU 驱动，通过 bootstrap.ps1 建环境并以 file: 路径安装到 desktop profile。默认 NPU 推理、CPU 跑 VAD 与 fbank，首次 bucket 编译约 2.5 分钟，适合 Core Ultra 用户将重计算移出 CPU。
Read Moredsh-voice-input：在 DSH Web 对话里本地离线语音转文字
dsh-voice-input 是 moazzamak 维护的 DSH 社区插件，面向 web profile 的 Web GUI 对话输入框。浏览器录音后把音频交给本机 harness，用 faster-whisper（默认 CPU）或 whisper.cpp/Vulkan（可选 GPU）离线识别，文本追加到草稿且不自动发送。支持 warm 模型、静音切分下的边说边出字，以及可选的 cleanup 去口语 filler。安装需 dsh plugin --profile web add github:moazzamak/dsh-voice-input 并运行包内 python/setup.py。适合重视隐私、在本机跑 Whisper 的 DSH 用户；使用前请核对源码与 MIT 许可证。
Read MoreOpenai Whisper Skill：用 Whisper CLI 本地语音转文字
Openai Whisper 是 SkillHub 上的 Agent Skill（@clawhub_steipete/openai-whisper，v1.0.0），指导 Agent 调用本机 Whisper CLI 完成语音转文字，标签 requires_api_key=false。文章说明其定位、核心能力、通过 skillhub install 安装到 Cursor 等 skills 目录的方法，以及上游提供的转写与翻译示例命令。适合 design-media 场景下处理本地音视频素材的开发者；需自行安装 whisper 并承担模型下载与算力成本。
Read More语音转文字：从 URL 拉取音频并用 Faster-Whisper 本地转录
语音转文字（@user_03f2ab96/yuyinzhuanxie）是 SkillHub 社区 Skill，从 URL 下载音频并用 Faster-Whisper 在本机转写，支持 YouTube、小宇宙及 yt-dlp 可提取的站点，无需云端 API。本文介绍其定位、核心能力、skillhub 安装命令与各 Agent skills 目录，并说明通过自然语言触发转写任务的典型用法与使用注意。
Read More讯飞听见语音转文字：在 Agent 里完成音频转写与会议纪要
介绍 SkillHub 技能「讯飞听见语音转文字」（@user_294a7000/xftj-transcribe，v1.0.4）：纯 Python 封装讯飞听见转写，支持新文件一键转写（上传、轮询、权益与支付、下载）、历史列表与结果下载、获取或创建 AI 会议纪要。说明 requires_api_key、安装提示与典型自然语言用法，适合在 Cursor 等环境中用对话完成录音转文字与纪要整理。
Read More抖音视频提取文案：本地 Playwright + Whisper 把分享链接转成文字
介绍 SkillHub 社区 Skill「抖音视频提取文案」（@user_6a36b82b/douyin-transcribe-lz，v1.1.1）。当用户提供 v.douyin.com、douyin.com、iesdouyin.com 链接或整段分享口令，并提出提取文案、视频转文字、语音转录或「这个视频说了什么」等意图时，Agent 应启用该 Skill。底层用 Playwright 绕过登录墙捕获视频，Whisper medium 本地转录中文，无需云端 API，requires_api_key=false。文中说明 skillhub CLI 安装命令、各客户端 skills 目录及典型对话示例。
Read More视频内容智能分析工具：用链接完成下载、转写与摘要
视频内容智能分析工具是 SkillHub 社区 Skill（v1.0.2，@user_09d1a78f/video-content-analyzer-plus），面向今日头条、抖音、B站、西瓜视频等链接，自动下载视频、用 faster-whisper 语音转文字，支持繁简转换与内容摘要。requires_api_key=false。安装方式为向 AI 发送 SkillHub 页面提示，按 skillhub.md 安装该坐标。适用于在 Agent 对话中分析、总结视频内容；需注意平台范围、环境与输出需人工校对。
Read Morezemanzhang809/dsh-stt-plugin: DeepSeek Harness Voice Input Plugin
The DeepSeek Harness plugin integrates the browser's Web Speech API to provide speech-to-text functionality in the chat interface. It supports microphone toggling, recognition language settings, single/continuous dictation modes, and auto-send, with all audio processed locally. Installation requires the environment to meet the prerequisites, and during use, clicking the toolbar button starts real-time voice input. It is fully supported on Chrome and Edge, partially supported on Safari, and not supported on Firefox.
Read MoreDeepSeek Harness: Minimal Voice Input Demo
The DeepSeek Harness Web UI Chinese Voice Input Plugin is designed to improve Chinese input efficiency in the Web interface. It relies on the browser's native speech recognition API. By clicking the microphone button on the right side of the input field, it converts speech into text in real time and appends it to the input field. After recognition, it does not automatically send the message, allowing users to edit it manually. Before use, the plugin must be installed via the terminal and the service must be restarted. Use cases include quick Chinese input and assisted note-taking. Note that it only supports the Web platform and the Chrome browser, and audio processing relies on browser services, so offline availability is not guaranteed.
Read Moredsh-voice-input-plugin: DeepSeek Harness Web Voice Input Plugin
`dsh-voice-input-plugin` is a zero-dependency plugin for DeepSeek Harness that uses the browser-native Web Speech API to perform local speech-to-text. The plugin requires no build step, provides a microphone button through the Web GUI toolbar, supports real-time word-by-word transcription and continued editing while speaking, and does not upload audio to the backend. Installation requires DSH version `>=0.1.0-rc.5`, and the Web Profile must be restarted for it to take effect. It currently only supports Chrome and Edge browsers, relies on an online speech recognition service, and the recognition language follows the browser's language settings.
Read Moredsh-asr-voice: Voice Input and Prompt Optimization
DSH Harness has launched the voice input plugin `dsh-asr-voice`, designed to address the disruption of thought continuity caused by voice and convert spoken language into clean text. Its core capabilities include a hybrid recognition architecture that supports both browser-native and multiple cloud-based ASR providers; a three-step configuration wizard that allows reuse of DSH credentials and automatically populates the BaseURL; and built-in privacy safeguards, with API keys stored only in DSH and audio recordings locally transcoded before upload. The plugin supports real-time voice conversations, half-duplex mode, keyboard shortcuts, and LLM optimization, eliminating the need for manual API key management and providing users with an efficient, secure, and closed-loop voice interaction experience.
Read MoreDeepSeek Harness Plugin: dsh-audio-input (Local Speech-to-Text on macOS)
`dsh-audio-input` is a web plugin for DeepSeek Harness (DSH) that adds a microphone button to the message editor. The plugin uses macOS’s native speech framework to run a local helper process for speech-to-text conversion, without requiring a cloud API key, ensuring that audio data never leaves the machine. Installation requires running commands on the server Mac and configuring a remote token. For remote voice input, the remote Mac must download and run the helper process. Configuration supports language settings and a remote mode toggle. To use the plugin, click the microphone button to record and insert text, while keeping in mind service restart requirements and system permission settings.
Read Moredsh-web-speech-input: Add a microphone button to the DeepSeek Harness Web input field
dsh-web-speech-input is a speech-to-text plugin for the DeepSeek Harness Web UI (maintained by leozou320-ai, v1.0.0, MIT license). Its core functionality is adding a microphone button to the left side of the prompt composer. Upon clicking, it uses the browser's Web Speech API to write real-time speech recognition results into the current draft. It supports interim and final recognition text, automatic language selection (Page Language -> Browser Language -> zh-CN fallback), clear status indicators, and accessibility support. The plugin never sends a prompt automatically; the submission action is always performed by the user. In terms of privacy, microphone permission is only requested after the user clicks. There are no server-side components, no audio storage, and no network requests are sent; audio processing depends on browser implementation. Installation Requirements
Read MoreUse Transcribe Skill to convert meeting recordings and podcasts into searchable text
### 正式翻译： This article introduces OpenAI's official Agent Skill `transcribe`: Based on `gpt-4o-mini-transcribe` and `gpt-4o-transcribe-diarize`, it converts audio and video files into text, with support for speaker diarization and known speaker prompts. It explains the installation directories and activation methods in Codex, Cursor and Claude Code, provides reproducible examples including fast transcription with the bundled CLI and `diarized_json` output, and lists precautions such as the 25MB file size limit, automatic chunking for long audio, and the fact that the diarize model does not support prompts.
Read More