YouTube Transcript MCP Server: Setup Guide for Claude and Other AI Tools
How the Model Context Protocol works, how to connect the GetYouTubeTranscript MCP server to Claude Desktop, Claude Code, Cursor, Windsurf, or ChatGPT, and how to troubleshoot the connection.
00:08:00 · SEP 13, 2026
Quick answer
What MCP actually is
The Model Context Protocol standardizes how an AI client discovers and calls external tools. Under the hood it's JSON-RPC 2.0 - the client asks the server what tools exist and their input schemas, then invokes them by name with structured arguments and gets a structured result back, all inside the same conversation turn.
MCP servers run over one of two transports: stdio, where the client launches the server as a local subprocess (typical for tools that need local filesystem access), or Streamable HTTP, where the server is remote and reachable over HTTPS by any number of clients at once - which is what this server uses. (An older HTTP+SSE transport exists in some documentation but is deprecated in favor of Streamable HTTP.) Practically, this means you never run anything locally to use this server - just point your client at the URL.
Available tools
| Tool | What it does | Cost |
|---|---|---|
get_youtube_transcript | Full transcript plus title, author, and thumbnail metadata | 1 credit |
search_youtube | Search YouTube for videos or channels, paginated | 1 credit |
get_channel_latest_videos | A channel's metadata plus its most recent uploads | Free |
search_channel_videos | Search within one channel for a query, paginated | 1 credit |
list_channel_videos | List every video a channel has uploaded, paginated | 1 credit |
list_playlist_videos | Get every video in a playlist, paginated | 1 credit |
The paginated tools (search, list-channel, list-playlist) all accept a continuation token from a previous response to fetch the next page, rather than an offset/limit pair.
Get an API key
Sign up from the dashboard and create an API key - 100 credits are included free, no card required. This key is the bearer token the MCP server authenticates with for most clients.
Add the server to Claude Desktop or Claude Code
Add this to your MCP client's config (e.g. claude_desktop_config.json, or via claude mcp add for Claude Code):
{
"mcpServers": {
"getyoutubetranscript": {
"url": "https://www.getyoutubetranscript.com/api/mcp",
"headers": {
"Authorization": "Bearer sk_live_..."
}
}
}
}Replace sk_live_... with your own API key. Restart the client fully and the 6 tools above appear automatically - most clients only read this config on startup, not on a hot reload.
Other clients: Cursor, Windsurf, and custom agents
Any client that supports a remote MCP server over Streamable HTTP with custom headers uses the same shape of config - just the surrounding file and field names differ (e.g. Cursor's .cursor/mcp.json, Windsurf's MCP settings panel). If a client only supports OAuth-based servers rather than a static bearer header, use the OAuth flow described next instead.
ChatGPT connectors and OAuth-only clients
For clients that authenticate MCP servers via OAuth rather than a pasted API key (for example, ChatGPT's custom connectors), this server also supports OAuth 2.1 with Dynamic Client Registration - point the client at the same server URL and it discovers the OAuth metadata and walks you through an authorization flow instead of asking for a bearer token directly.
Ask it to use a tool
Once connected, natural prompts are enough - the model decides which tool to call:
- "Get the transcript for youtube.com/watch?v=dQw4w9WgXcQ and summarize it in 3 bullets."
- "Search YouTube for 'Next.js server actions' and list the top 5 results."
- "List every video @mkbhd has uploaded and find ones mentioning 'battery'."
Troubleshooting
Building a custom agent instead of using a chat client? The same server works with any MCP-compatible client over Streamable HTTP. Full REST endpoint reference, credits, and rate limits are in the API docs.
MCP server FAQs
What is an MCP server, and why would I use one for YouTube transcripts?
MCP (Model Context Protocol) lets an AI assistant like Claude call external tools directly during a conversation. Instead of pasting a transcript into the chat, you ask Claude to fetch it - it calls the tool, gets the transcript back, and can summarize, translate, or search it in the same turn.
Is MCP the same as a plain REST API wrapper?
No. MCP standardizes tool discovery and invocation across any compliant client - the same server works with Claude Desktop, Claude Code, Cursor, Windsurf, or a custom agent without writing client-specific glue code. A REST API still requires you (or the model) to know the exact endpoint shape; MCP tools are self-describing, so the client lists them and their input schemas automatically.
What transport does this MCP server use?
Streamable HTTP - the current transport for remote, multi-client MCP servers (the older HTTP+SSE transport is deprecated). Under the hood, MCP messages are JSON-RPC 2.0, and the transport layer just handles getting those messages to and from the server; you don't need to know JSON-RPC to use it, only the config below.
Do I need an API key to use the MCP server?
Yes for most tools. Requests are authenticated with a bearer token - the same API key you'd use for the REST API. Sign up for 100 free credits from the dashboard, no card required. This server also supports OAuth 2.1 (Dynamic Client Registration) for clients like ChatGPT connectors that authenticate via an OAuth flow instead of a static header.
Does this cost credits the same way as the REST API?
Yes - each successful metered tool call costs 1 credit, same accounting as the REST API. get_channel_latest_videos is free and never charges a credit. Failed calls are never charged.
The tools aren't showing up after I added the config - what do I check?
Restart the client completely (not just reload the window) - most clients only read MCP server config on startup. Then check that the bearer token is valid (an expired or malformed key returns 401, not a missing-tools symptom) and that the URL has no typo - a wrong path returns 404 before authentication is even checked.
Is it safe to put my API key in a config file?
Treat an MCP config file like any other credential file - don't commit it to a public repo, and prefer your client's secrets/environment-variable support over a hardcoded key if it offers one. The connection itself runs over HTTPS, so the token isn't exposed in transit.
Related