MCP YouTube Transcript
An MCP server retrieving transcripts of YouTube videos
作者提供的说明(英文原文)
Give your agent marketing superpowers: research, renders, delivery, paid campaigns, measurement.
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。
Authorization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 …”。
这个地址需要在请求的 header 中放入密钥,而添加连接器的菜单无法设置。请改用 Claude Code、Cursor 或 VS Code。
打开 Claude Desktop → Settings → Developer → Edit Config,把下面的内容粘贴到 claude_desktop_config.json
{
"mcpServers": {
"layers": {
"command": "npx",
"args": [
"-y",
"@layers/cli",
"mcp"
]
}
}
}在终端中运行这条命令
claude mcp add --transport http layers https://mcp.layers.com/mcp --header 'Authorization: your-value'claude mcp add layers -- npx -y @layers/cli 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": {
"layers": {
"url": "https://mcp.layers.com/mcp",
"headers": {
"Authorization": "your-value"
}
}
}
}粘贴到项目中的 .cursor/mcp.json,或 ~/.cursor/mcp.json
{
"mcpServers": {
"layers": {
"command": "npx",
"args": [
"-y",
"@layers/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": {
"layers": {
"type": "http",
"url": "https://mcp.layers.com/mcp",
"headers": {
"Authorization": "your-value"
}
}
}
}粘贴到项目中的 .vscode/mcp.json
{
"servers": {
"layers": {
"type": "stdio",
"command": "npx",
"args": [
"-y",
"@layers/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。
这个地址需要在请求的 header 中放入密钥,而添加连接器的菜单无法设置。请改用 Claude Code、Cursor 或 VS Code。
你的选择通过 Cookie 保存在这台设备上,无需账户。
Authorization 机密 必填 HTTP header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。
我们还没有读取这个服务器的工具列表;我们会轮流检查各个端点。
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 layers 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
尚未检查;我们会轮流检查各个端点
尚未检查;我们会轮流检查各个端点
尚未检查;我们会在检查端点时读取工具列表
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:Authorization
两种方式都可以:安装在你的电脑上,或者连接云端版本。
GitHub 仓库最后更新于–,最新版本(2.1.16)于2个月前进入注册中心。
支持 MCP 的应用,例如 Claude Desktop、Claude Code、Cursor 和 VS Code。它通过 URL 连接,所以接受远程 MCP 服务器的应用也可以使用,例如开启开发者模式的 ChatGPT。
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。
Give your agent marketing superpowers. Layers Marketing MCP is the endpoint your coding agent calls to run growth for your product: onboarding, the competitive map, what to post, the plan, the persona, the render, delivery, the review, paid campaigns, Layers-operated accounts, and halt on one call. Every tool carries its price in credits, and credits come out of your organization's balance when the tool runs.
This repository holds the public description of that endpoint and the server.json that MCP directories read. The server is a hosted service.
https://mcp.layers.com/mcpapplication/json or text/event-stream, and Mcp-Session-Id rides both directions.Authorization: Bearer https://mcp.layers.com/healthThe configuration to prefer is layers mcp as a stdio server. Your client spawns it, it forwards every JSON-RPC message with the session from your OS keyring, and it is the only thing that reads that credential.
{
"mcpServers": { "layers": { "command": "layers", "args": ["mcp"] } } }
Install the CLI with npm install -g @layers/cli, brew install layers/tap/layers, or curl -fsSL https://layers.com/install.sh | sh. Source: layers/cli.
Paste this into Claude Code, Codex, Cursor, or any coding agent working in your repository. The agent runs both commands. The browser sign-in is the only step that needs a person.
Set up Layers Growth for me so you can plan, render and deliver content for this product from here.
npm install -g @layers/cli.layers setup from this repository. It opens my browser once for sign-in and waits for me; it writes the .mcp.json entry and the Layers skill into the repo and stores the session in my OS credential store, so nothing secret lands in the repo..agents/skills/layers/SKILL.md, then tell me what to post first.Let me know when it is ready.
/plugin marketplace add layers/mcp
/plugin install layers@layers
The plugin carries the layers MCP entry, the Layers skill, and three commands. /layers:setup walks the paste above. /layers:post takes one opportunity through plan, render and delivery as a draft, quoting the cost before every charged call. /layers:status reads layers://account, layers://workspace and layers://jobs and summarizes them.
The plugin configures the client. The CLI holds the session and .layers/project.json binds the project, so install @layers/cli and run layers setup in the repository as well. In a repository that has both, Claude Code lists the plugin's entry as plugin:layers:layers and the one layers setup wrote as layers. They reach the same server, and one of them is enough. Codex and Cursor read the same .mcp.json entry and the same skill that layers setup writes, so there is no separate plugin for them.
One credit balance per organization, and the same operation costs the same on every plan. Credits are bought in packs from one of two shelves: pay as you go is $0 a month and prices credits at two cents each, and a paid plan buys the same pack sizes cheaper. Packs add to the same balance and do not expire. Every charged tool takes quoteOnly, which returns the exact maximum credits before anything runs, so your agent can read the price before it spends. Plenty of the catalog charges nothing at all: reading any layers:// resource, reading the standing opportunity queue, reading stored experiment results, planning a paid campaign, quoting a managed fleet, storing one of your own brand assets, creating a persona, minting a tracking link, installing measurement, and halt. Ad spend is billed by the ad platform against your own ad account, and a Layers-operated account is $150 per account per month from your cash wallet; neither one consumes credits.
Bind a product from the repository it is already in and start its first landscape build; keep the map of accounts it is measured against; read TikTok live for creators, their posts, a topic, a hashtag or the sound chart; turn one opportunity into an editable plan; create a persona and render a reaction reel, a slideshow, a still ad or a remix of a public post; deliver an artifact to a platform draft inbox or post it directly; register a post you published yourself and score it against that account's own normal; install the Layers SDK in your own repository and mint the tracking link that resolves an install back to a click; plan, draft, launch and adjust paid campaigns under exact caps and stop conditions; order and release Layers-operated TikTok accounts; and stop everything in scope on one call.
The live tool list is tools/list on the endpoint. The full generated inventory with prices is https://growth.layers.com/llms-full.txt.
send_feedback sends a message to the founders, who read every one. A bug, a price you think is wrong, a capability you looked for and did not find, or something that worked. It is free, and it works for any member of the workspace, including one whose organization has run out of credits.
The contents of this repository are MIT licensed. See LICENSE. The Layers MCP server is a hosted service, and its source is not in this repository.