MCP YouTube Transcript
An MCP server retrieving transcripts of YouTube videos
作者提供的说明(英文原文)
AI trend analysis agents: what's trending on Google, TikTok, YouTube and Reddit, and if it lasts
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 …”。
适用于网页版、桌面版和手机版 Claude:
https://trendsagents.com/mcp在终端中运行这条命令
claude mcp add --transport http trend-analysis https://trendsagents.com/mcp粘贴到项目中的 .cursor/mcp.json,或 ~/.cursor/mcp.json
{
"mcpServers": {
"trend-analysis": {
"url": "https://trendsagents.com/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": {
"trend-analysis": {
"type": "http",
"url": "https://trendsagents.com/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。
ChatGPT 只能连接互联网上的服务器:
https://trendsagents.com/mcp菜单名称可能因套餐和版本不同而略有差异。
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。
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6日
注册名称已通过域名 trendsagents.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
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。
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。
它在作者的云端运行。通过互联网连接,无需安装,但你发送的内容会经过作者的服务器。
GitHub 仓库最后更新于–,最新版本(1.0.0)于16天前进入注册中心。
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。
AI agents for trend analysis and trend forecasting. Each agent takes a live board (Google Trends, TikTok, YouTube, Reddit, Amazon and 22 more), then checks growth and history to tell a spike from a trend that lasts.
[Website](https://trendsagents.com/?utm_source=github&utm_medium=readme&utm_campaign=trend-analysis-agents) · [Get started free](https://trendsagents.com/account?utm_source=github&utm_medium=readme&utm_campaign=trend-analysis-agents) · Docs · Pricing · All agents
You: Something is trending on YouTube. Should I make a video about it?
Claude: (YouTube agent: reads YouTube Trending, then YouTube search growth for the title) Film, wait, or skip, with the growth numbers behind the call...
Point any MCP client at https://trendsagents.com/mcp. In Claude and ChatGPT you sign in; no key to handle.
Building your own agent loop? Use the same tools over MCP with a key, or call the REST API:
claude mcp add --transport http trendsagents https://trendsagents.com/mcp
Free key and the agent prompts: trendsagents.com/account. Setup per client: trendsagents.com/docs.
Tip: start with "Using Trends Agents" so your AI uses these tools instead of a web search.
Each agent starts from a live board, then checks growth or history to answer one question.
| Agent | Answers |
|---|---|
| Google Trends agent | Spike versus lasting demand |
| Google Shopping agent | Shopping demand versus browsing |
| YouTube trending agent | Film, wait, or skip |
| TikTok hashtags agent | Meme versus staying tag |
| Reddit Hot agent | Thread spike or real demand |
| Amazon Best Sellers agent | Chart hit versus searched product |
| Wikipedia pageviews agent | News burst versus a page people keep opening |
| Google News agent | Headline versus Google News search interest |
| Google Play agent | Chart position versus installs |
| App Store agent | Top Free versus Top Paid |
| npm downloads agent | Package growth versus a rival |
| Steam Charts agent | Weekend spike versus a game people still play |
| PyPI stats agent | Python package growth, e.g. pandas versus polars |
| X trending agent | Trending topic versus search growth |
| Spotify podcasts agent | Chart versus people searching the show |
| GitHub trending agent | Trending repo versus package downloads |
| IMDb MOVIEmeter agent | Popularity versus searches for the title |
| Open Library agent | Trending book versus search and page views |
| Most visited websites agent | Traffic rank versus branded search |
| Substack agent | Leaderboard versus searches for the newsletter |
Every agent runs the same three steps with these tools:
get_top_trends: one of 27 live boards, current leaders or the biggest climbers.get_growth and get_time_series: short window against long window on 14 keyword sources.get_context: what people are posting about it on News, YouTube, TikTok, Instagram, X, Reddit, Bluesky, Hacker News, Polymarket, GitHub.| Tool | Cost |
|---|---|
get_time_series | 1 request per call |
get_growth | 1 request per call, every window included |
get_top_trends | MCP: 1 request per 10 rows, at most 10 per call. REST: 1 request per call |
get_context | MCP: 1 request per source, at most 5 per call. REST: one source per call, 1 request |
get_top_posts | 1 request per call (Business plan) |
get_usage | Free |
Sources an agent can check growth on
source | Example keyword |
|---|---|
Google Search | electric vehicles |
Google Images | electric vehicles |
Google News | electric vehicles |
Google Shopping | running shoes |
YouTube | stanley cup |
Wikipedia | Bitcoin |
TikTok | booktok |
Amazon | air fryer |
Pinterest | matcha |
npm | langchain |
Python | pandas |
Steam | Elden Ring |
App Downloads | com.openai.chatgpt |
App Rankings | com.openai.chatgpt |
Every board, option and keyword format: trendsagents.com/data-catalog.
Each agent step is one request (a board read, a growth check, one context source). The free plan's 100 requests a month cover about one three-step run a day. Scheduled agents fit Starter ($19, 1,000) or Pro ($49, 5,000); Business ($199, 25,000) adds hashtag top posts. Run out mid-month? Credit packs never expire: $15 for 600, $49 for 2,500, $199 for 12,000 credits. Details: trendsagents.com/pricing. Start free.
Trend vs fad: how do I tell? A fad spikes and falls back within weeks. A trend keeps growing over months. Each agent checks both: the short window (7D, 1M) against the long one (12M, 5Y).
What is AI trend analysis? An AI assistant reads live trend boards and growth data for you, then explains what is rising and whether it lasts. These agents give Claude, Cursor and ChatGPT that data through MCP.
Which platforms are covered? Google, Google Images, Google News, Google Shopping, YouTube, Wikipedia, TikTok, Amazon, Pinterest, npm, PyPI, Steam, Google Play, App Store, GitHub, Google Trends, IMDb, Open Library, Polymarket, Reddit, Web, Spotify, Substack, X, Instagram, Bluesky and Hacker News.
Is it free? Yes. The free plan has 100 requests a month with every source. No credit card.
MIT for the contents of this repository. Use of the hosted service follows the terms of service.