💯 YouTube Transcript + YouTube Search MCP for AI Agents

What 💯 YouTube Transcript + YouTube Search MCP for AI Agents does

💯 YouTube Transcript + YouTube Search MCP for AI Agents MCP server exposes a hosted set of 12 tools for working with YouTube content through MCP. Its capabilities cover transcript retrieval, video and channel metadata, video search, channel browsing, searches limited to a channel, playlist extraction, and monitoring for new uploads through RSS.

The server is intended for agent workflows that need more than a transcript from one known video. An agent can search for relevant videos, identify content from a particular channel, retrieve transcripts, and use the returned material for summarization or comparison. The README gives a workflow that searches YouTube, searches a channel’s videos, and then retrieves transcripts for selected results.

How it works

The MCP endpoint is hosted remotely at https://transcriptapi.com/mcp, so the client communicates with TranscriptAPI rather than starting a local stdio process. MCP-compatible clients can connect through the endpoint using OAuth or a bearer API key. OAuth authorization is handled by the client and can be initiated in a browser; API keys are available from the TranscriptAPI dashboard.

The 💯 YouTube Transcript + YouTube Search MCP for AI Agents MCP server can also be installed as an Agent Plugins 1.0.0 package. The repository includes plugin and MCP manifests plus a bundled YouTube skill that describes when to use the tools. VS Code and Cursor installation paths are documented, and Claude supports adding the endpoint as a custom connector. The repository also provides a Cursor deeplink for adding the remote server without cloning it.

Setup and configuration

You need a TranscriptAPI account. The README documents two authentication paths: an API key or OAuth through supported clients such as Claude and ChatGPT. The first account tier includes 100 credits and does not require a card, according to the setup instructions.

For Cursor, configure a remote MCP server named transcript-api with the URL https://transcriptapi.com/mcp; an API key can be sent in the Authorization header. For Claude, add a custom connector using the same URL and authorize it in the browser. Claude Code can add the server with its HTTP transport command. VS Code can install the repository as a plugin from its source URL or register a local clone in settings.json.

No environment variables are specified as required. Credentials are not embedded in the repository; authorization is supplied by the client or through the configured API key.

Tools and capabilities

The documented tool groups support:

• Fetching a full transcript for a YouTube video, including timestamps.
• Searching YouTube videos by topic or query.
• Reading video and channel metadata.
• Browsing videos from a channel.
• Searching within a channel’s videos.
• Extracting playlist contents.
• Checking for a channel’s latest uploads through RSS.

These functions place the project in the 🔎 Search & Data Extraction category. Reach for it when the agent must discover YouTube material and then extract source text for analysis in the same workflow.

Limitations and notes

Access depends on TranscriptAPI authentication and available account credits. The remote endpoint avoids local installation, but it also means the MCP client must support the documented remote connection or plugin flow. The README lists Claude, ChatGPT, Cursor, VS Code, Claude Code, and other clients, while installation details vary by client. The repository is licensed under MIT.

The 💯 YouTube Transcript + YouTube Search MCP for AI Agents MCP server is focused on YouTube data retrieval and extraction; the provided material does not document general web search, video downloading, or publishing actions.

Getting started with this 💯 YouTube Transcript + YouTube Search MCP for AI Agents MCP server

Always refer to the official documentation for the most accurate and up-to-date information.