MCP YouTube Transcript
An MCP server retrieving transcripts of YouTube videos
作者提供的说明(英文原文)
IOCs (URLs, domains, IPs, hashes) shared by the infosec community on X/Twitter. No auth, CC0.
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://mcp.tweetfeed.live/在终端中运行这条命令
claude mcp add --transport http tweetfeed https://mcp.tweetfeed.live/粘贴到项目中的 .cursor/mcp.json,或 ~/.cursor/mcp.json
{
"mcpServers": {
"tweetfeed": {
"url": "https://mcp.tweetfeed.live/"
}
}
}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": {
"tweetfeed": {
"type": "http",
"url": "https://mcp.tweetfeed.live/"
}
}
}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://mcp.tweetfeed.live/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8日
注册名称已通过 GitHub 账户 0xDanielLopez 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。
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(0.1.0)于4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。
Model Context Protocol (MCP) server for [tweetfeed.live](https://tweetfeed.live).
Exposes the public IOC feed (URLs, domains, IPs, SHA256/MD5 hashes shared by the infosec community on Twitter/X) as MCP tools so AI agents can query threat intel programmatically.
https://mcp.tweetfeed.live/ (HTTP JSON-RPC 2.0, POST)| Name | Purpose |
|---|---|
query_iocs | Query IOCs by time window (today/week/month) with optional user, tag, and type filters. |
check_url | Check whether a specific URL appears in the feed. |
check_ip | Check whether an IPv4/IPv6 address appears in the feed. |
check_hash | Check whether an MD5 or SHA-256 hash appears in the feed (type auto-detected). |
list_recent_iocs | List IOCs added since a given date, with optional type/tag filters. |
get_tag_info | Window aggregates plus recent IOCs for a tag (leading # optional). |
get_trending | Top tags and IOC-type distribution for a window (today/week/month/year). |
enrich_ioc | Exact lookup of an IOC (auto-detected type: url/domain/ip/md5/sha256) over the past 365 days, with AI/external-corroboration/net/domain-registration (reg, including the domain's age in days at first report and a newly_registered flag) context and campaign membership when available, plus an archive of history older than 365 days when it exists (can accompany a live match); falls back to a 30-day substring scan on a miss. |
get_campaigns | AI-clustered campaign groupings from the last 30 days, with optional brand and min-confidence filters. |
get_campaign_iocs | Full IOC membership (header + rows) of one campaign by id, with optional type filter and limit. |
get_trends | 31-day IOC trend analytics: daily volume by type, top moving tags week-over-week, most-abused TLDs, new vs recurring ratio, producer concentration. |
get_feed_status | Live pipeline health: freshness verdict per artifact (stale, age_seconds) and source coverage (which hashtag/account feeds delivered rows in the last 24h, which account feeds are dead). |
search | ChatGPT connector interface: find document ids (ioc: , tag: , campaign: ) for an IOC value, tag, campaign id or free text. |
fetch | ChatGPT connector interface: full document ({id, title, text, url, metadata}) for an id returned by search. |
Besides the 14 tools, the server exposes the feed as MCP resources (resources/list, resources/templates/list, resources/read): 7 static documents plus 2 URI templates, each returning the raw JSON of the matching api.tweetfeed.live document. Tools are the filtered, size-bounded path; resources are for clients that want a whole document (a status panel, a whole-window blocklist sync).
| URI | Content | ||||||
|---|---|---|---|---|---|---|---|
tweetfeed://status | Pipeline health plus per-source coverage (mirrors /v1/status, ~13 KB, refreshed every 15 min). | ||||||
tweetfeed://counts | Aggregate IOC counts per window/type/tag (mirrors /v1/counts, ~6 KB). | ||||||
tweetfeed://trends | 31-day trend analytics (mirrors /v1/trends, ~4 KB, daily). | ||||||
tweetfeed://campaigns | AI-clustered campaigns of the trailing 30 days (mirrors /v1/campaigns, ~300 KB, daily). | ||||||
tweetfeed://iocs/today | Every IOC row shared today (mirrors /v1/today, 50-250 KB by the end of the UTC day). | ||||||
tweetfeed://iocs/week | Last 7 days (mirrors /v1/week, 1-1.5 MB). | ||||||
tweetfeed://iocs/month | Last 30 days (mirrors /v1/month, 2-3 MB, API caps at 10,000 rows). | ||||||
tweetfeed://iocs/{time}/{type} | Template - time: today\ | week\ | month, type: url\ | domain\ | ip\ | sha256\ | md5 (mirrors /v1/{time}/{type}). |
tweetfeed://campaigns/{id} | Template - id = tfc- plus 12 hex chars (mirrors /v1/campaigns/{id}). |
An unknown URI or an expired campaign id returns JSON-RPC error -32002 "Resource not found".
Besides the 14 tools and the resources above, the server also speaks MCP prompts (prompts/list, prompts/get, params name and arguments as strings): ready-made recipes that chain several tool calls into one instruction sequence for the model to run and report back on. Where tools are chosen by the model and resources are raw documents, prompts are recipes the user invokes directly - they show up as slash commands in Claude Desktop and Claude Code.
| Name | Arguments | Purpose | |
|---|---|---|---|
triage_ioc | ioc (required), context (optional) | Step-by-step triage of one IOC (URL/domain/IP/MD5/SHA-256): enrichment, campaign membership, verdict and recommended action. | |
daily_brief | window (optional, today default \ | week) | Short brief: feed health, volume and mix, moving tags, active high-confidence campaigns, abused TLDs. |
campaign_summary | campaign_id (required, tfc- plus 12 hex) | Hand-off summary of one campaign: context, targeting, ATT&CK techniques, IOC breakdown, infrastructure, block-ready lists, CSV/STIX links. |
{
"mcpServers": { "tweetfeed": { "url": "https://mcp.tweetfeed.live/" } } }
Or from the Claude Code CLI:
claude mcp add tweetfeed https://mcp.tweetfeed.live/
ChatGPT (Settings > Connectors > Create, or a Deep research custom connector) accepts https://mcp.tweetfeed.live/ with no authentication. The server exposes the required search and fetch tools alongside the others, and every result carries a tweetfeed.live URL so ChatGPT can cite it.
curl -sX POST https://mcp.tweetfeed.live/ \
-H 'content-type: application/json' \ -d '{"jsonrpc":"2.0","method":"tools/list","id":1}' | jq .
curl -sX POST https://mcp.tweetfeed.live/ \ -H 'content-type: application/json' \ -d '{"jsonrpc":"2.0","method":"tools/call","id":2, "params":{"name":"query_iocs", "arguments":{"time":"today","tag":"phishing","type":"url","limit":5}}}' | jq .
curl -sX POST https://mcp.tweetfeed.live/ \ -H 'content-type: application/json' \ -d '{"jsonrpc":"2.0","method":"resources/read","id":3, "params":{"uri":"tweetfeed://status"}}' | jq .
curl -sX POST https://mcp.tweetfeed.live/ \ -H 'content-type: application/json' \ -d '{"jsonrpc":"2.0","method":"prompts/get","id":4, "params":{"name":"triage_ioc","arguments":{"ioc":"example.com"}}}' | jq .
npm install
npm run dev # wrangler dev on http://localhost:8787 MCP_URL=http://localhost:8787 npm test
npm run deploy # wrangler deploy (routes mcp.tweetfeed.live/*)
MCP_URL=https://mcp.tweetfeed.live npm test