whisper-windows-mcp

whisper-windows-mcp

A Windows-native MCP server that lets Claude Desktop transcribe audio files locally using whisper.cpp, with no internet connection required.

Category
访问服务器

README

whisper-windows-mcp

A Windows-native MCP (Model Context Protocol) server that lets Claude Desktop transcribe audio and video files locally using whisper.cpp — no internet connection required, no data sent to the cloud.

Why does this exist? The popular whisper-mcp package was built for macOS and assumes a Unix environment. It does not work on Windows. This package was written specifically for Windows users who want the same local transcription functionality in Claude Desktop.


What you can do with it

Once installed, you can say things like this directly in Claude Desktop:

  • "Transcribe C:\Users\Me\Downloads\meeting.mp3"
  • "Transcribe C:\Users\Me\Videos\interview.mp4" — video files work directly, no conversion needed
  • "Transcribe this recording and summarise the key points"
  • "Transcribe with timestamps so I can find specific moments"
  • "Generate subtitles for C:\Users\Me\Videos\lecture.mp4"
  • "Transcribe all files in C:\Users\Me\Videos\clips\ one at a time"

Everything runs on your own machine. No audio ever leaves your computer.


Requirements

Requirement Purpose Download
Node.js 18+ Runs the MCP server nodejs.org
whisper.cpp The transcription engine Latest release
A Whisper model file The AI model See Step 2 below
FFmpeg Video file support ffmpeg.org

FFmpeg is optional if you only use MP3/WAV files, but required for MP4, MKV, AVI, MOV and other video formats.


Step 1 — Install whisper.cpp

  1. Go to the whisper.cpp latest release
  2. Download the file named whisper-bin-x64.zip (look for win and x64 in the filename)
  3. Extract the ZIP and move the contents to C:\whisper\Release\ — create this folder if it doesn't exist

✅ You should now have C:\whisper\Release\whisper-cli.exe

Why this path? You can install whisper.cpp anywhere, but C:\whisper\Release\ matches the default config below and means less to edit later.


Step 2 — Download a Whisper model

Click a link below to download directly:

Model Download Size Speed Best for
tiny.en Download 75 MB Very fast Quick tests
base.en Download 142 MB Fast Recommended starting point
small.en Download 466 MB Moderate Better accuracy
medium.en Download 1.5 GB Slow High accuracy
large-v3 Download 2.9 GB Very slow Maximum accuracy

Save the downloaded .bin file to C:\whisper\models\ — create this folder if it doesn't exist.

✅ You should now have something like C:\whisper\models\ggml-base.en.bin


Step 3 — Install Node.js

If you don't already have Node.js:

  1. Go to nodejs.org and download the Windows Installer (.msi) — choose the LTS version
  2. Run the installer and accept all defaults

✅ Verify: open Command Prompt and run node --version — you should see something like v20.x.x


Step 4 — Install FFmpeg (recommended)

FFmpeg is required for video files (MP4, MKV, AVI, MOV, etc.).

  1. Go to ffmpeg.org/download.html and download a Windows build
  2. Extract and move the bin folder contents (or the whole folder) somewhere permanent, e.g. C:\ffmpeg\bin\
  3. Add C:\ffmpeg\bin to your system PATH:
    • Press Win + S → search Environment Variables → open it
    • Under User Variables, select PathEditNew
    • Add C:\ffmpeg\bin → click OK on all dialogs

✅ Verify: open a new Command Prompt and run ffmpeg -version


Step 5 — Configure Claude Desktop

  1. Open Claude Desktop → Settings → Developer → Edit Config
  2. Add the following (or merge the mcpServers block if you have other servers):
{
  "mcpServers": {
    "whisper": {
      "command": "npx",
      "args": ["-y", "whisper-windows-mcp"],
      "env": {
        "WHISPER_CLI_PATH": "C:\\whisper\\Release\\whisper-cli.exe",
        "WHISPER_MODEL": "C:\\whisper\\models\\ggml-base.en.bin"
      }
    }
  }
}

⚠️ Path format: In the JSON config, all backslashes must be doubled (\\). This is a JSON requirement. When typing paths into Claude in chat, use normal single backslashes.

  1. If you downloaded a different model, update ggml-base.en.bin to match your filename
  2. Save the file, fully quit Claude Desktop, and reopen it
  3. Go to Settings → Developer — you should see whisper with a green running badge

Step 6 — Test it

In Claude Desktop, type:

"Can you check your whisper config?"

Claude will verify that everything is found correctly. Then try:

"Please transcribe C:\Users\YourName\Downloads\recording.mp3"


Available tools

Tool What it does
transcribe_audio Transcribe a single file — audio or video
transcribe_batch Transcribe all files in a folder, one at a time with preview
generate_subtitles Generate an .srt subtitle file next to the source
check_config Verify all paths and FFmpeg are working

Output formats

Format What you get Ask Claude...
text (default) Plain transcript, no timestamps "Transcribe this file"
timestamps Transcript with [00:00:00 --> 00:00:05] time codes "Transcribe with timestamps"
json Structured data "Transcribe as JSON"
srt Subtitle file saved next to source "Generate subtitles for..."

Supported file formats

Audio (native): MP3, WAV

Audio (via FFmpeg): M4A, FLAC, OGG

Video (via FFmpeg): MP4, MKV, AVI, MOV, WebM, FLV, WMV, M4V

H.264 and H.265/HEVC video both work. FFmpeg must be installed and in your PATH for any video or non-MP3 audio format.


Batch transcription

To transcribe multiple files in a folder interactively:

"Transcribe all files in C:\Users\Me\Videos\clips"

Claude will list all detected files (with checkmarks on already-completed ones), then process them one at a time, showing you a preview of each transcript before moving to the next.

For large unattended overnight batches, use whisper-cli directly from the command line — see TROUBLESHOOTING.md for the syntax. This is more reliable than running through Claude for very large jobs.


Transcription speed

Whisper runs on CPU by default. Rough estimates for a 1-hour recording:

Model Approximate time (CPU)
tiny.en 5–10 minutes
base.en 10–20 minutes
small.en 20–35 minutes
medium.en 35–60 minutes
large-v3 60–120 minutes

You can increase thread count by adding "WHISPER_THREADS": "12" to the env block in your config (replace 12 with however many threads you want — up to your CPU's logical core count).

GPU acceleration for AMD (ROCm) and NVIDIA (CUDA) on Windows is planned for a future update.


Converting video to audio (optional)

whisper-windows-mcp handles video files automatically via FFmpeg, so manual conversion is no longer required. However, if you want a smaller audio-only file for any reason, VLC makes it easy:

  1. Open VLC → Media → Convert / Save
  2. Click Add, select your video, then Convert / Save
  3. Under Profile, choose Audio - MP3
  4. Set a destination filename and click Start

Full config example

{
  "preferences": {
    "coworkWebSearchEnabled": true
  },
  "mcpServers": {
    "whisper": {
      "command": "npx",
      "args": ["-y", "whisper-windows-mcp"],
      "env": {
        "WHISPER_CLI_PATH": "C:\\whisper\\Release\\whisper-cli.exe",
        "WHISPER_MODEL": "C:\\whisper\\models\\ggml-base.en.bin",
        "WHISPER_THREADS": "8",
        "FFMPEG_PATH": "ffmpeg"
      }
    }
  }
}

Config file location:

C:\Users\YourUsername\AppData\Roaming\Claude\claude_desktop_config.json

The AppData folder is hidden by default. To show it: File Explorer → View → Show → Hidden items


Tested on

  • Windows 10 Pro (10.0.19045)
  • Windows 11 — untested, feedback welcome via Issues

Troubleshooting

See TROUBLESHOOTING.md for detailed solutions including how to run large overnight batch jobs from the command line.

Quick checklist:

  • [ ] Config paths use double backslashes (C:\\whisper\\...)
  • [ ] whisper-cli.exe exists at the path specified
  • [ ] The model .bin file exists at the path specified
  • [ ] Claude Desktop was fully restarted after editing the config
  • [ ] Whisper shows running in Settings → Developer
  • [ ] FFmpeg is in PATH if using video files

Roadmap

  • [ ] AMD GPU acceleration (ROCm)
  • [ ] NVIDIA GPU acceleration (CUDA)
  • [ ] Speaker diarization (automatic A/B speaker identification)
  • [ ] Translation to English from other languages
  • [ ] Unattended background batch processing

Support this project

If this tool saved you time and you'd like to support continued development:

  • Star this repo — it helps others find it
  • 💬 Open an issue if you find a bug or have a feature request
  • 💖 SponsorGitHub Sponsors | Ko-fi | Patreon

License

MIT — free to use, modify, and distribute.


Contributing

Pull requests welcome. GPU acceleration for AMD or NVIDIA especially appreciated. Windows 11 feedback welcome via Issues.

推荐服务器

Baidu Map

Baidu Map

百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。

官方
精选
JavaScript
Playwright MCP Server

Playwright MCP Server

一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。

官方
精选
TypeScript
Magic Component Platform (MCP)

Magic Component Platform (MCP)

一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。

官方
精选
本地
TypeScript
Audiense Insights MCP Server

Audiense Insights MCP Server

通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。

官方
精选
本地
TypeScript
VeyraX

VeyraX

一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。

官方
精选
本地
graphlit-mcp-server

graphlit-mcp-server

模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。

官方
精选
TypeScript
Kagi MCP Server

Kagi MCP Server

一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。

官方
精选
Python
e2b-mcp-server

e2b-mcp-server

使用 MCP 通过 e2b 运行代码。

官方
精选
Neon MCP Server

Neon MCP Server

用于与 Neon 管理 API 和数据库交互的 MCP 服务器

官方
精选
Exa MCP Server

Exa MCP Server

模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。

官方
精选