MCP YouTube Transcript
An MCP server retrieving transcripts of YouTube videos
作者提供的说明(英文原文)
Manage Transistor.fm podcasts: episodes, analytics, transcripts, and webhooks.
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。
TRANSISTOR_API_KEY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": {
"transistor": {
"command": "npx",
"args": [
"-y",
"transistor-mcp"
],
"env": {
"TRANSISTOR_API_KEY": "your-value"
}
}
}
}在终端中运行这条命令
claude mcp add transistor --env TRANSISTOR_API_KEY=your-value -- npx -y transistor-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": {
"transistor": {
"command": "npx",
"args": [
"-y",
"transistor-mcp"
],
"env": {
"TRANSISTOR_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": {
"transistor": {
"type": "stdio",
"command": "npx",
"args": [
"-y",
"transistor-mcp"
],
"env": {
"TRANSISTOR_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。
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 conorbronsdon 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。
你需要填写:TRANSISTOR_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(0.4.0)于3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。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。
An MCP server for the Transistor.fm API. Manage podcasts, episodes, analytics, transcripts, and webhooks from your AI agent.
[!IMPORTANT]
Transistor now ships an official MCP server: [mcp.transistor.fm](https://mcp.transistor.fm). It's remote (no install), authenticates with OAuth instead of an API key, and is maintained by the Transistor team as the basis for their upcoming API v2. For most people it's the better choice — start there.
With the official server available, this community server is no longer actively maintained. It still works, and it remains useful if you specifically want a local, API-key-based server (for example, in headless or automated environments where an OAuth browser flow is awkward). Bug-fix PRs are welcome; feature requests are better directed at the official server.
This MCP server provides tools to interact with the Transistor.fm API, allowing you to manage podcasts, episodes, and view analytics. The data shown in the demo above is sample data, not real account values.
Fork notice: The original server was built by Guido X Jansen (gxjansen/Transistor-MCP). The full-API-parity pass (all documented params, response trimming, search), the
get_download_summaryandcompare_episodesanalytics tools, ISO date handling, and transcript support were contributed here and have since been merged upstream. This repository is a packaged, npm-published build (transistor-mcp). The MIT license and original copyright are preserved in LICENSE.
Built by Conor Bronsdon for the Chain of Thought podcast production workflow, where it handled episode creation, transcript uploads, and analytics pulls before the official server shipped. Conor hosts Chain of Thought, a show about AI infrastructure and how practitioners actually build with it. More tools for creators live in ai-tools-for-creators. Find Conor on X at @ConorBronsdon.
Companion tools:
Prefer the official server? No install needed. Add it to Claude Code with:
claude mcp add --transport http transistor https://mcp.transistor.fm
Then run /mcp to sign in to Transistor. Other MCP clients that support remote servers with OAuth can point at https://mcp.transistor.fm directly.
To run this community server instead, add it to your MCP settings configuration file with your Transistor API key:
{
"mcpServers": { "transistor": { "command": "npx", "args": ["-y", "transistor-mcp"], "env": { "TRANSISTOR_API_KEY": "your-api-key-here" } } } }
Or run from a local clone (after npm install && npm run build):
{
"mcpServers": { "transistor": { "command": "node", "args": ["path/to/Transistor-MCP/build/index.js"], "env": { "TRANSISTOR_API_KEY": "your-api-key-here" } } } }
Get details of the authenticated user account.
{
// No parameters needed }
Get a pre-signed URL for uploading an audio file. Use this before creating an episode with a local audio file.
{
"filename": string // Required: Name of the audio file to upload }
Response includes:
List all shows in your Transistor.fm account, ordered by updated date (newest first). Returns a paginated list with 10 items per page.
{
"page": number, // Optional, defaults to 0 (first page) "per": number, // Optional, defaults to 10 items per page "private": boolean, // Optional: filter for private shows "query": string // Optional: search query }
Note: All parameters are optional. Calling this endpoint without parameters will return the first page of shows.
List episodes for a specific show.
{
"show_id": string, // Required "page": number, // Optional, defaults to 0 "per": number, // Optional, defaults to 10 "query": string, // Optional: search query "status": string, // Optional: "published", "draft", or "scheduled" "order": string // Optional: "desc" (newest first) or "asc" (oldest first), defaults to "desc" }
Get detailed information about a specific episode.
{
"episode_id": string, // Required "include": string[], // Optional: array of related resources to include "fields": { // Optional: sparse fieldsets "episode": string[], // Fields to include for episode "show": string[] // Fields to include for show } }
Get analytics for a show or specific episode. Defaults to the last 14 days if no dates are provided.
{
"show_id": string, // Required "episode_id": string, // Optional: include for episode-specific analytics "start_date": string, // Optional: format "dd-mm-yyyy", required if end_date is provided "end_date": string // Optional: format "dd-mm-yyyy", required if start_date is provided }
Create a new episode.
{
"show_id": string, // Required "title": string, // Required "audio_url": string, // Required "summary": string, // Optional "description": string, // Optional: may contain HTML "transcript_text": string, // Optional: full episode transcript "author": string, // Optional "explicit": boolean, // Optional "image_url": string, // Optional: episode artwork "keywords": string, // Optional: comma-separated list "number": number, // Optional: episode number "season_number": number, // Optional "type": string, // Optional: "full", "trailer", or "bonus" "alternate_url": string, // Optional: override share_url "video_url": string, // Optional: YouTube URL "email_notifications": boolean, // Optional: override show setting "increment_number": boolean // Optional: auto-set next episode number }
Update an existing episode.
{
"episode_id": string, // Required "title": string, // Optional "summary": string, // Optional "description": string, // Optional: may contain HTML "transcript_text": string, // Optional: full episode transcript "author": string, // Optional "explicit": boolean, // Optional "image_url": string, // Optional: episode artwork "keywords": string, // Optional: comma-separated list "number": number, // Optional: episode number "season_number": number, // Optional "type": string, // Optional: "full", "trailer", or "bonus" "alternate_url": string, // Optional: override share_url "video_url": string, // Optional: YouTube URL "email_notifications": boolean // Optional: override show setting }
Get analytics for all episodes of a show. Defaults to the last 7 days if no dates are provided.
{
"show_id": string, // Required "start_date": string, // Optional: format "dd-mm-yyyy", required if end_date is provided "end_date": string // Optional: format "dd-mm-yyyy", required if start_date is provided }
List all webhooks for a show.
{
"show_id": string // Required }
Subscribe to a webhook for a show.
{
"event_name": string, // Required: e.g., "episode_created" "show_id": string, // Required "url": string // Required: URL to receive webhook events }
Unsubscribe from a webhook.
{
"webhook_id": string // Required }
Tool calls that fail against the Transistor.fm API return a status-specific, human-readable message via isError: true rather than a generic Axios error. Internally, TransistorApiClient maps every HTTP failure — authentication (401/403), rate limiting (429), validation (400), not found (404), and server errors (5xx) — to a typed error class in a single response interceptor (see src/errors.ts), so the message you see always names the failure mode and, for auth errors, points at the TRANSISTOR_API_KEY environment variable.
fieldsresource_typeinclude[] to fetch related resources in a single request["resource_name"]List shows:
// List first page of shows (default behavior)
const result = await use_mcp_tool({ server_name: "transistor", tool_name: "list_shows", arguments: {} });
// List shows with pagination and filtering const resultWithParams = await use_mcp_tool({ server_name: "transistor", tool_name: "list_shows", arguments: { page: 1, per: 20, private: true, query: "podcast" } });
Get episode details:
const result = await use_mcp_tool({
server_name: "transistor", tool_name: "get_episode", arguments: {
只显示了开头部分;其余内容请在仓库中阅读