MCP Youtube Transcribe

MCP开发工具

提供方 JackHP

开发工具 未标注 无需鉴权 官方维护

Enables fetching transcripts from YouTube videos by searching for videos based on text queries and retrieving either official transcripts or generating them using AI-powered transcription with Whisper when official transcripts are unavailable.

一句话选型结论

按资料字段整理

MCP Youtube Transcribe 属于开发工具类 MCP 服务器,传输方式未标注,无需鉴权,没有现成安装命令,需按官方文档接入。接入前先确认权限范围;涉及写入、删除或付款的操作,建议先在测试环境验证。

分类开发工具 传输方式未标注 鉴权无需鉴权 部署方式本地部署
GitHub 星标 未标注 仓库热度,无公开仓库时显示未标注
工具数 未标注 暴露给模型的工具数量
资源 / 提示词 0 / 0 资源与提示词模板数量
维护方 官方维护 核验等级 仅资料核对

工具与能力

开发工具
数量来自来源目录的标注,具体工具清单以官方文档为准
工具数量未标注暴露给模型可调用的工具个数
资源数量0可读取的资源条目
提示词数量0内置提示词模板数量

能力标签

安装与配置

未标注 无需鉴权
上游未给出传输方式
  1. 1
    准备运行环境

    stdio 服务在本地起进程,需要先装好 Node.js 等运行时,再按安装命令拉取依赖。

  2. 2
    安装服务

    复制上面的安装命令执行;命令里带占位符的位置,换成自己的密钥或路径。

  3. 3
    写入客户端配置

    在客户端的 MCP 配置里填好启动命令和参数,保存后完全退出客户端再打开,让新配置生效。

  4. 4
    验证是否可用

    连接成功后随便调一个只读工具,确认能返回结果,再接入正式流程。

典型使用场景

3 类
按分类与标签整理的常见用法,具体以官方文档为准
代码与仓库操作

在对话里直接查代码、提 issue、走 PR,省掉来回切工具。

本地命令与脚本

把构建、测试、格式化这类重复命令交给模型执行。

接口联调与测试

让模型调你的接口、跑一遍用例,快速定位问题。

选择建议

选之前先看这几条
按条目字段给出的判断,不替你决定用不用

值得选的地方

  • 资料标注无需鉴权,装好就能连。

先想清楚的地方

  • stdio 方式要求本地装好运行时(如 Node.js),换机器要重新装一遍。
  • 没有现成安装命令,需要按官方文档自己填启动命令和参数。
  • 资料没标注工具数量,实际能力要装好后确认。
  • 版本更新与行为细节以官方仓库为准。

更适合

  • 需要在开发工具场景里让模型直接调用外部能力的人。
  • 想先快速试一下、能接受在本地或自有环境跑服务的人。

不太适合

  • 完全不想在本地装环境的人。
  • 需要厂商级 SLA 和技术支持的生产环境。

同类对比

6 个同类
同分类里挑最接近的几条,字段以各自详情页为准
MCP 服务器工具数传输方式部署方式安装命令最后核验
MCP Youtube Transcribe当前 未标注 未标注 本地部署 无 2026-05-17
Git 838e2ec1 12 stdio · 本地进程 本地部署 有 2026-05-17
Markitdown MCP 2c653a5f 1 stdio · 本地进程 本地部署 有 2026-05-17
JavaGuide 未标注 stdio · 本地进程 本地部署 有 2026-04-11
awesome-claude-skills 未标注 stdio · 本地进程 本地部署 有 2026-04-26
Context7 831acf40 未标注 stdio · 本地进程 本地部署 有 2026-05-17

服务器档案

开发工具
字段来自项目仓库与公开索引,中文说明由本站标注
提供方JackHP上游标注的维护方 JackHP
分类开发工具本站按上游分类展示
连接方式未标注上游未给出传输方式
鉴权方式无需鉴权远程服务建议用最小权限凭据
维护方官方维护核验等级 仅资料核对
包名 / 运行时未标注 · 未标注本地安装时用到的包标识与运行时
部署类型本地部署上游给出的部署形态
最近核验2026-05-17上游最后核验时间

详细介绍

9.3 KB
来自项目仓库与上游目录的原文,未做删改

MCP-YouTube-Transcribe

An MCP server that provides a tool to fetch transcripts from YouTube videos. It first attempts to
retrieve a pre-existing, official transcript. If one is not available, it downloads the video's audio and uses OpenAI's
Whisper model for local AI-powered transcription.

This project is designed to be a simple, self-contained tool that can be easily integrated into any system capable of
communicating with an MCP server.

Features

  • YouTube Video Search: Finds the most relevant YouTube video based on a text query.
  • Official Transcript Priority: Intelligently fetches manually created or auto-generated YouTube transcripts first
    for speed and accuracy.
  • Fast AI-Powered Transcription: Uses whisper.cpp (if available) for blazing fast transcription. Falls back to OpenAI's Python Whisper
    tiny model if whisper.cpp is not installed.
  • MCP Server Interface: Exposes the transcription functionality as a simple tool (get_youtube_transcript) via the
    lightweight model context protocol.

Requirements

  • Python 3.12+
  • uv: A fast Python package installer and resolver. You will need to install
    uv on your system first.
  • FFmpeg: Must be installed and available in your system's PATH. Required for audio processing.
  • whisper.cpp (Highly recommended): MCP-YouTube-Transcribe will first try to use whisper.cpp for lightning-fast local transcription and only fall back to Python Whisper if the executable is not found.
  • macOS: brew install whisper-cpp
  • Linux: Build from source following the whisper.cpp installation guide
  • Windows: Build from source or grab a pre-built binary from the releases page

After installation, make sure the whisper-cli (or whisper-cpp on older versions) command is in your PATH.

Finally, download a Whisper model. The tiny model offers the best speed-to-quality ratio for most use-cases:bash mkdir -p models curl -L -o models/ggml-tiny.bin \ https://huggingface.co/ggerganov/whisper.cpp/resolve/main/ggml-tiny.bin

Place additional models in the same models/ folder if you wish to experiment.

Installation with uv

Using uv is recommended as it's extremely fast and handles both environment creation and package installation
seamlessly.

  1. Clone the repository:bash git clone https://github.com/<your-username>/YouTubeTranscriber.git cd YouTubeTranscriber

  2. Create and activate a virtual environment:
    This command creates a .venv folder in your project directory and activates it. uv will automatically use this
    environment for all subsequent commands.bash uv venv

  3. Install the project and its dependencies:
    This command reads the pyproject.toml file and installs all required libraries into the virtual environment.bash uv sync

Usage

Running the MCP Server

Once installed, you can start the server by running the mcp_server.py script. The server will listen for JSON-RPC
requests on stdin and send responses to stdout.```bash
python mcp_server.py


The server will log its activity to a file named `mcp_server.log` in the project's root directory.

## Connecting to Gemini CLI on Windows

You can connect this MCP server to the Google Gemini CLI to use the function as a native tool directly from your
terminal. These instructions are for a **Windows** environment.

### Step 1: Create a Startup Script `run_server.bat`

The Gemini CLI needs a single, reliable command to start your server. A batch script is the perfect way to handle this
on Windows, as it ensures the correct virtual environment and Python interpreter are used.

1. Create a new file named in the root of your project directory. `run_server.bat`
2. Copy and paste the following content into the file:``` batch
    @echo off
    REM This ensures the script's directory is the current directory
    cd /d "%~dp0"

    REM --- IMPORTANT ---
    REM Replace the path below with the ABSOLUTE path to your project's venv python.exe
    set PYTHON_EXE="C:\Users\jackp\.pyenv\pyenv-win\versions\3.12.10\python3.12.exe"

    echo --- Starting MCP Server using %PYTHON_EXE% ---
    %PYTHON_EXE% mcp_server.py

    pause

This script activates the virtual environment in your project and then runs the server, ensuring all the correct
dependencies are available..venv

Step 2: Configure the Gemini CLI

Now, you need to tell the Gemini CLI how to find and run your new server.

  1. Locate your Gemini CLI config.json file. On Windows, this is typically found at:
    C:\Users\\.gemini\config.json
  2. Open the config.json file in a text editor. Add the following entry to the mcpServers object. If mcpServers
    doesn't exist, create it as shown below.``` json
    {
    "mcpServers": {
    "MCP-YouTube-Transcribe": {
    "command": "C:\Windows\System32\cmd.exe",
    "args": [
    "/c",
    "<path-to-your-project>\run_server.bat"
    ],
    "cwd": "<path-to-your-project>"
    }
    }
    }

3. **Crucially, you must replace both instances of ``** with the full, absolute path to where you
   cloned the `YouTubeTranscriber` repository.

**Example:** If your project is located at `C:\dev\YouTubeTranscriber`, the entry would look like this:``` json
    {
      &quot;mcpServers&quot;: {
        &quot;MCP-YouTube-Transcribe&quot;: {
          &quot;command&quot;: &quot;C:\\Windows\\System32\\cmd.exe&quot;,
          &quot;args&quot;: [
            &quot;/c&quot;,
            &quot;C:\\dev\\YouTubeTranscriber\\run_server.bat&quot;
          ],
          &quot;cwd&quot;: &quot;C:\\dev\\YouTubeTranscriber&quot;
        }
      }
    }

Note: JSON requires backslashes to be escaped, so you must use double backslashes (\\) in your paths.

Step 3: Verify the Connection

After saving the config.json file, you can verify that Gemini CLI recognizes and can use your new tool.

Run Gemini CLI and press ctrl+t

You should see MCP-YouTube-Transcribe listed as an available tool.

Connecting to Gemini CLI on Mac/Unix

You can also connect this MCP server to the Google Gemini CLI on Mac or other Unix-like systems. The process is similar to Windows but uses a shell script instead of a batch file.

Step 1: Prepare the Startup Script

The repository already includes a run_server.sh script. Just make it executable:```bash
chmod +x run_server.sh


### Step 2: Configure the Gemini CLI

1. Locate your Gemini CLI `config.json` file. On Mac/Unix systems, this is typically found at:
   `~/.gemini/config.json`

2. Open the `config.json` file in a text editor. Add the following entry to the `mcpServers` object. If `mcpServers`
   doesn&#x27;t exist, create it as shown below:```json
{
  &quot;mcpServers&quot;: {
    &quot;MCP-YouTube-Transcribe&quot;: {
      &quot;command&quot;: &quot;/path/to/your/project/run_server.sh&quot;,
      &quot;cwd&quot;: &quot;/path/to/your/project&quot;
    }
  }
}
  1. Replace both instances of /path/to/your/project with the absolute path to where you cloned the repository.

Example: If your project is located at /Users/username/MCP-YouTube-Transcribe, the entry would look like this:```json
{
"mcpServers": {
"MCP-YouTube-Transcribe": {
"command": "/Users/username/MCP-YouTube-Transcribe/run_server.sh",
"cwd": "/Users/username/MCP-YouTube-Transcribe"
}
}
}


### Step 3: Verify the Connection

After saving the `config.json` file, you can verify that Gemini CLI recognizes and can use your new tool.

Run Gemini CLI and press ctrl+t

You should see `MCP-YouTube-Transcribe` listed as an available tool.

### MCP Client Example

You can interact with the server using any client that supports the MCP protocol over stdio. The server exposes one
primary tool: `get_youtube_transcript`.

Here is an example of a `call_tool` request to get a transcript for the query &quot;What is an API? by MuleSoft&quot;.

**Request:**```json
{
  &quot;jsonrpc&quot;: &quot;2.0&quot;,
  &quot;id&quot;: 1,
  &quot;method&quot;: &quot;call_tool&quot;,
  &quot;params&quot;: {
    &quot;name&quot;: &quot;get_youtube_transcript&quot;,
    &quot;arguments&quot;: {
      &quot;query&quot;: &quot;What is an API? by MuleSoft&quot;,
      &quot;force_whisper&quot;: false
    }
  }
}
  • query: The search term for the YouTube video.
  • force_whisper: (Optional) A boolean that, if true, skips the check for an official transcript and generates one
    directly with Whisper. Defaults to false.

Testing

This project includes a test suite to verify its functionality.

  • Core Function Test (simple.py): This script tests the server's handler functions directly without needing to run
    a separate server process. It's the quickest way to check if the core logic is working.bash python simple.py

  • Full Server Test (test_mcp.py): This script starts the MCP server as a subprocess and sends it live JSON-RPC
    requests, providing an end-to-end test of the server's functionality.bash python test_mcp.py

Configuration

  • Logging: Server activity is logged to mcp_server.log.
  • Audio Cache: When Whisper is used, downloaded audio files are temporarily stored in testing/audio_cache/. You
    may wish to change this path in youtube_tool.py for a production environment.

Contributing

Contributions are welcome! If you'd like to improve the YouTube Transcriber, please feel free to fork the repository and
submit a pull request.

Please read our CONTRIBUTING.md for details on our code of conduct and the process for submitting pull requests to us.

License

This project is licensed under the MIT License - see the LICENSE file for details.

常见问题

7 问
答案由本站依据条目字段整理,不引用上游原文
MCP Youtube Transcribe 怎么安装?

资料没有给出安装命令,请到官方仓库或文档按说明接入;这类条目多数是生态应用,或尚未发布可安装包。

MCP Youtube Transcribe 怎么在客户端里配置?

它的传输方式未标注,在客户端 MCP 配置里按官方文档填写启动命令和参数,保存后重启客户端生效。

MCP Youtube Transcribe 收费吗?

本站不代收费用,也没有标注具体价格。是否收费取决于服务本身和背后的接口额度,正式使用前按官方说明确认。

MCP Youtube Transcribe 需要 API Key 吗?

资料标注为无需认证,装好就能连;如果实际调用时报鉴权错误,按官方文档补上凭据。

MCP Youtube Transcribe 能提供哪些能力?

资料显示能力标签包括 开发工具、命令行工具、自动化。具体工具清单以官方文档为准。

MCP Youtube Transcribe 需要什么运行环境?

属于本地进程,需要在本机装好运行环境。

MCP Youtube Transcribe 是本地还是远程?

它是本地进程(stdio),由客户端在本机拉起。

热门专题

按标签挑服务器
全部专题 →

服务器信息以作者仓库与官方文档为准;分类、传输方式与鉴权方式的中文说明为中文整理,最近核验 2026-05-17。