MCP YouTube Transcript
An MCP server retrieving transcripts of YouTube videos
作者提供的说明(英文原文)
Read-only Google Search Console analytics, URL inspection and sitemaps via SearchDock CLI.
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。
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": {
"searchdock-cli": {
"command": "npx",
"args": [
"-y",
"searchdock-cli",
"mcp"
]
}
}
}在终端中运行这条命令
claude mcp add searchdock-cli -- npx -y searchdock-cli mcp粘贴到项目中的 .cursor/mcp.json,或 ~/.cursor/mcp.json
{
"mcpServers": {
"searchdock-cli": {
"command": "npx",
"args": [
"-y",
"searchdock-cli",
"mcp"
]
}
}
}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": {
"searchdock-cli": {
"type": "stdio",
"command": "npx",
"args": [
"-y",
"searchdock-cli",
"mcp"
]
}
}
}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。
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7日
注册名称已通过 GitHub 账户 xplormityai 验证,该账户拥有此仓库
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。
注册中心没有列出必需的密钥,不过有些服务器会在首次使用时要求你登录。
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(0.1.3)于26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。
Read Google Search Console analytics, inspect indexed URLs, and monitor sitemaps from your terminal or AI agent.
SearchDock is open source, runs locally, and requests read-only Google access. Use the same core through npm, Python, MCP, a Codex or Claude Code plugin, or a standalone agent skill. Bing Webmaster Tools support is planned.
Copy this prompt into your coding agent:
Set up SearchDock CLI for this project using
https://github.com/xplormityai/searchdock-cli. Read its README and skills/searchdock-setup/SKILL.md. Check Node.js 22+, reuse an existing installation, or install the CLI from npm. Install searchdock-setup and searchdock skills for the agent I am using, scoped to this project, and connect the stdio MCP server if this agent supports it. Keep existing agent configuration. For Google access, ask only for the local path to my Desktop OAuth client JSON and guide me through read-only browser consent. Never print credentials or tokens. Let me choose the Search Console property, then verify the connection with a metrics request. Tell me exactly which steps succeeded and what still needs my input.
Node.js 22 or later is required for every distribution.
# npm
npm install --global searchdock-cli searchdock --help
npx --yes searchdock-cli@0.1.3 --help
pipx install searchdock-cli
The Python package is a launcher containing the exact same JavaScript runtime. It has no separate Google implementation. If you install both packages, use python -m searchdock_cli to select the Python launcher explicitly.
Install both skills for Codex in the current project:
npx --yes skills add xplormityai/searchdock-cli --skill searchdock-setup searchdock --agent codex --copy --yes
Replace codex with claude-code, cursor, opencode, or your supported agent. Add --global only for installation across projects. Skills do not install the CLI or authorize Google access. Plugin, MCP and manual setup.
To run from source:
git clone https://github.com/xplormityai/searchdock-cli.git
cd searchdock-cli npm ci npm run build node dist/cli.cjs --help
Create a Google OAuth Desktop app client in your own Google Cloud project, enable the Search Console API, and download the client JSON. See Google setup for the steps and testing-mode limits. No shared developer credentials are included.
searchdock auth login --client-file /path/to/desktop-client.json
searchdock sites list searchdock sites use sc-domain:example.com searchdock auth status
Google opens a local browser consent flow. SearchDock uses PKCE, validates the callback state, and listens only on 127.0.0.1. It requests webmasters.readonly and keeps the refresh token in your operating system's per-user configuration directory, outside this repository. Credentials are stored as permission-protected local JSON, not encrypted by SearchDock. See privacy.
# Search performance and top queries
searchdock analytics --days 28 searchdock analytics --dimensions page --days 28 --json searchdock analytics --dimensions date --start 2026-09-01 --end 2026-09-30
searchdock analytics --brand "Your Brand" --json
searchdock sitemaps searchdock inspect https://example.com/
searchdock report --urls https://example.com/,https://example.com/about --json searchdock report --output ./private-report.json
searchdock auth login --client-file /path/to/client.json --profile work searchdock sites use https://example.com/ --profile work searchdock analytics --profile work --site https://example.com/ --json
searchdock doctor searchdock auth logout
--json writes machine-readable results to stdout and errors to stderr with a nonzero exit status. --output saves JSON to a new file and refuses to overwrite an existing one. Keep private reports out of Git yourself. CSV is not yet supported. For URLs containing commas, use the MCP urls array or the JavaScript API for batch inspection.
Report windows default to 28 days ending three days before today in Pacific Time, to allow for Search Console processing. You can specify exact dates. Finalized data may still lag. Pagination supports up to 100,000 requested rows, but Google's API returns top rows and omits some anonymized query data. Missing rows do not prove zero demand; zero impressions make average position meaningless.
SearchDock provides five read-only MCP tools: searchdock_sites, searchdock_analytics, searchdock_sitemaps, searchdock_inspect, and searchdock_report. Authenticate locally first; MCP tools do not start consent or change account settings.
Add this stdio server to any compatible MCP client:
{
"mcpServers": { "searchdock": { "command": "npx", "args": ["--yes", "searchdock-cli@0.1.3", "mcp"] } } }
For a source checkout, set command to node and args to ["/absolute/path/searchdock-cli/dist/cli.cjs", "mcp"]. Windows clients that launch npx.cmd through a shell may need command: "cmd" and args: ["/c", "npx", "--yes", "searchdock-cli@0.1.3", "mcp"]. Use your client's Windows configuration guidance.
Plugins use the same pinned npm version:
# Codex CLI
codex plugin marketplace add xplormityai/searchdock-cli codex plugin add searchdock@searchdock
In Claude Code:
/plugin install searchdock --marketplace xplormityai/searchdock-cli --scope project
This one-step command requires Claude Code 2.1.275+. Earlier versions can use /plugin marketplace add xplormityai/searchdock-cli, then /plugin install searchdock@searchdock --scope project.
For standalone skills, copy both skills/searchdock and skills/searchdock-setup into your agent's skills directory. Codex uses ~/.agents/skills/; Claude Code uses ~/.claude/skills/. Install the CLI separately. Both skills are included in the plugin. The repository includes portable Agent Plugins, Codex and Claude manifests; local stdio plugins are not automatically listed in hosted plugin directories.
You can also install the skill with the open skills installer:
npx --yes skills add xplormityai/searchdock-cli --skill searchdock-setup searchdock --agent codex --copy --yes
SearchDock is published in the official MCP Registry under io.github.xplormityai/searchdock-cli. Release workflows update its metadata using GitHub OIDC.
Both SearchDock and setup skills are published on ClawHub. Its skill copies use MIT-0, as required by that registry; this repository and its packages remain MIT licensed.
The local MCPB bundle is published on Smithery. It requires Node.js 22+ and a locally authenticated Google profile; it includes no credentials. python scripts/build_mcpb.py packages a built dist/cli.cjs. Release and discovery automation.
| Available | Requires Search Console UI or an export |
|---|---|
| Clicks, impressions, CTR, average position | Request indexing |
| Top queries/pages and date/device/country breakdowns | Test a live URL |
| Exact brand-query performance | Full Page indexing exclusions table |
| Sitemap processing status | Manual actions and security issues |
| Google's indexed version of specific URLs | Guaranteed current ranking or indexing |
SearchDock's report flags returned sitemap errors/warnings, pending processing and non-PASS verdicts for the URLs you inspect. It is not a complete technical SEO audit, and it does not provide invented SEO/AEO/GEO scores.
import { selectedContext } from 'searchdock-cli';
const { adapter, site } = await selectedContext({ profile: 'default' }); const result = await adapter.analytics(site, { days: 28, dimensions: ['page'] }); console.log(result.rows);
The provider registry declares capabilities separately. Bing can be added as an adapter without changing the CLI, MCP or Python distribution model. Provider-specific differences must remain explicit. See architecture.
npm ci
npm run build npm test npm run check python -m pip install build twine npm run release:artifacts python -m twine check dist/releases/searchdock_cli-0.1.3-py3-none-any.whl dist/releases/searchdock_cli-0.1.3.tar.gz
Release instructions cover npm/PyPI accounts, OIDC trusted publishing, versioning and GitHub release artifacts. Package creation alone does not publish anything.
Built by Xplormity Collective. MIT license.