MCP YouTube Transcript
An MCP server retrieving transcripts of YouTube videos
作者提供的说明(英文原文)
Run SEO Content Machine tasks from an MCP client: scraping, crawls, Google Maps, Workspace pages.
这三项永远不会合并成一个分数:意见、测量和人气是不同的东西。
我们还没有为它写通俗的介绍。下面是我们从注册中心读取到的事实。
选择你使用的 AI 应用,然后复制设置。
SCM_API_URL, SCM_API_KEY标为机密的值是你账户的钥匙。千万不要分享给别人,也不要粘贴到聊天里。
你需要
你用的是哪个应用?
在上方选择你的应用。我们会记住它,下次按钮会显示“安装到 …”。
打开 Claude Desktop → Settings → Developer → Edit Config,把下面的内容粘贴到 claude_desktop_config.json
{
"mcpServers": {
"scm": {
"command": "npx",
"args": [
"-y",
"seo-content-machine-mcp"
],
"env": {
"SCM_API_URL": "your-value",
"SCM_API_KEY": "your-value"
}
}
}
}在终端中运行这条命令
claude mcp add scm --env SCM_API_URL=your-value --env SCM_API_KEY=your-value -- npx -y seo-content-machine-mcp运行前请先填写空着的值。
粘贴到项目中的 .cursor/mcp.json,或 ~/.cursor/mcp.json
{
"mcpServers": {
"scm": {
"command": "npx",
"args": [
"-y",
"seo-content-machine-mcp"
],
"env": {
"SCM_API_URL": "your-value",
"SCM_API_KEY": "your-value"
}
}
}
}浏览器会先询问是否打开应用,然后应用会显示设置并请你确认。密钥的值会留空,由你自己填写。
粘贴到项目中的 .vscode/mcp.json
{
"servers": {
"scm": {
"type": "stdio",
"command": "npx",
"args": [
"-y",
"seo-content-machine-mcp"
],
"env": {
"SCM_API_URL": "your-value",
"SCM_API_KEY": "your-value"
}
}
}
}浏览器会先询问是否打开应用,然后应用会显示设置并请你确认。密钥的值会留空,由你自己填写。
这个服务器在你的电脑上运行,而 ChatGPT 只能连接互联网上的服务器。请选择其他应用。
你的选择通过 Cookie 保存在这台设备上,无需账户。
SCM_API_URL env这些设置是根据作者在注册中心的条目生成的。如果 README 的写法不同,请以 README 为准。
它在你的电脑上运行;安装后,工具列表会显示在你的 AI 应用中。
我们检查的是任何人都能看到的信息;我们没有阅读代码。不适用的检查项不计入总数。 · 检查于 2026年11月6日
注册名称已通过域名 seocontentmachine.com 验证(注册中心会核实域名所有权)
我们还没有从 GitHub 读取这个仓库
我们还没有从 GitHub 读取这个仓库
它在你的电脑上运行,没有互联网端点
在你的电脑上运行,改用本地设置
它在你的电脑上运行;我们只能读取互联网端点的工具列表
与注册中心里这个分类中所有活跃的服务器比较。只有当发布者拥有该仓库时,星标数才计入。
它在 1 项适用的基础检查中通过了 1 项。我们没有阅读代码。在添加密钥或令牌之前,请先看看这个服务器要求哪些权限,并使用只具备所需权限的密钥。
你需要填写:SCM_API_KEY
它在你的电脑上运行。你的 AI 应用会在需要时启动它,数据不会经过作者的服务器(除非它自己调用在线服务)。
GitHub 仓库最后更新于–,最新版本(1.1.1)于4天前进入注册中心。
支持 MCP 的应用,例如 Claude Desktop、Claude Code、Cursor 和 VS Code。你的电脑上还需要有 Node.js。
同一分类中有:MCP YouTube Transcript, Google Analytics MCP, BibiGPT YouTube Video Summarizer
同一分类中其他作者的服务器,按星标数排序。
如何看卡片:★ GitHub 星标i三大支柱:评价 · 检查 · 使用情况i更新 = 最新版本发布到注册中心的时间i
An MCP server retrieving transcripts of YouTube videos
GA4 MCP server for AI agents: schema discovery, server-side aggregation, analysis-ready defaults.
AI video summarizer: summarize and transcribe YouTube, Bilibili, TikTok and podcast videos.
按原文从作者的仓库复制。PopNx MCP 没有审查或背书其内容。HTML、图片和脚本已被移除。链接会离开本站。
Connects Claude Code, Codex, Claude Desktop or any MCP client to SEO Content Machine (SCM), the desktop SEO workspace.
Your assistant can then create and run SCM tasks and read their results: article scraping from search results, site crawls, search result lists, Google Maps listings, YouTube subtitles, archive.org downloads and SEO Workspace pages. The work runs inside the SCM app on your machine, with its own browser and task queue.
Claude Code:
claude mcp add scm -- npx -y seo-content-machine-mcp
Any other MCP client, in its config file:
{
"mcpServers": { "scm": { "command": "npx", "args": ["-y", "seo-content-machine-mcp"] } } }
If you set an API key in SCM (Settings > App > Api secret key), pass it as SCM_API_KEY. If SCM's API runs somewhere other than http://localhost:8008, set SCM_API_URL.
| Tool | What it does |
|---|---|
task_list, task_get, task_results | List SCM tasks, read one with its log, read the result rows of its last run, which point to its output files |
task_run, task_cancel, task_delete | Start (or start over with fromStart), stop or delete a task |
page_pipeline_upsert | Create or change a Page Pipeline: AI pages from page templates, or articles scraped from search results, for a list of keywords |
dynamic_pages_create | Extract values from rendered pages into CSV with CSS selectors |
static_pages_create | Fetch pages without a browser and save each as a cleaned article |
site_crawler_create | Crawl a site and save the links it finds |
url_finder_create | Search a list of queries and save every result to one CSV of query, rank, url and title |
youtube_subtitles_create | Save the subtitles of YouTube videos, from video URLs or keyword searches |
web_archives_create | Download the newest archive.org copy of each page on a domain |
local_listings_create | Scrape Google Maps listings, such as "plumbers in denver" |
seo_workspace_create | Create an empty SEO Workspace |
workspace_query | Read an SEO Workspace: its pages, site plan, brief, documents, crawl and Search Console data |
workspace_write | Add or update pages, the site plan, the brief or documents in an SEO Workspace, with undo in the app |
workspace_create_pages | Hand SEO Workspace pages to a Page Pipeline, which researches, writes and reviews each one as a draft |
workspace_documents_edit | Replace exact text in one SEO Workspace document |
web_search, browser_open, cdp_open | Search the web, read a page, or drive a real browser |
render_pdf | Print an HTML document to a PDF file |
SCM serves MCP over HTTP at /mcp inside the app. This package is a small stdio relay: the client talks to it over stdin and stdout, and it forwards each message to the app. You can start SCM before or after your MCP client. While SCM is closed the server lists no tools, and it tells the client to list them again once SCM starts. Tool calls made while SCM is closed return an error that says to start it.