How to Summarize YouTube Videos in Cursor

Technical knowledge lives on YouTube in a format that’s hard to work with: you have to watch it linearly, you can’t search it, and you can’t paste it into your editor. But most videos have transcripts, and once you have the text, an AI agent can work with it like any other document. With AgentPatch connected, Cursor can pull a transcript and summarize it on the spot.

Why This Matters

Framework tutorials, conference talks, recorded architecture discussions, API walkthroughs. These are all common reference material during development. Watching a 40-minute video to find the three minutes that matter is a real time cost.

AgentPatch gives Cursor access to the YouTube Transcript tool. You provide a URL, it returns the full transcript with timestamps, and Cursor can summarize it, extract specific claims, or use the content to inform what you’re building. The video becomes searchable text inside your editor session.

Setup

The AgentPatch CLI is designed for AI agents to use via shell access. Install it, and your agent can discover and invoke any tool on the marketplace.

Install (zero dependencies, Python 3.10+):

pip install agentpatch

Set your API key:

export AGENTPATCH_API_KEY=your_api_key

Example commands your agent will use:

ap search "web search"
ap run google-search --input '{"query": "test"}'

Get your API key from the AgentPatch dashboard.

  1. Open Cursor Settings (Cmd+Shift+J on Mac, Ctrl+Shift+J on Windows/Linux).
  2. Navigate to the MCP section in the sidebar.
  3. Click Add new global MCP server.
  4. Cursor opens ~/.cursor/mcp.json. Add the AgentPatch server config:
{
  "mcpServers": {
    "agentpatch": {
      "url": "https://agentpatch.ai/mcp",
      "headers": {
        "Authorization": "Bearer YOUR_API_KEY"
      }
    }
  }
}
  1. Replace YOUR_API_KEY with your actual key from the AgentPatch dashboard.
  2. Save and restart Cursor. It discovers all AgentPatch tools automatically.

Example

You’re implementing an authentication flow and found a recorded walkthrough of the API. In Cursor’s chat:

“Summarize this video and find where they explain the OAuth flow: https://www.youtube.com/watch?v=xyz789”

Cursor fetches the transcript through AgentPatch, returns a summary, and points you to the section covering authentication with timestamps. You stay in the editor. The agent did the research.

For a broader use case, say you want the key points from a long talk:

“Summarize this video for me: https://www.youtube.com/watch?v=EKBIni-eB1A”

Cursor fetches the transcript and returns a structured summary:

“Lazy Potatoes” — Growing Potatoes With No Hilling or Top Dressing

The grower planted potatoes in a 10ft x 4ft raised bed using a “set it and forget it” approach. No hilling, no mid-season fertilizer. Plants grew aggressively but flopped without hilling, and leaves turned pale from insufficient feeding during tuber development.

Final harvest: ~60 lbs from 40 sq ft. He estimates hilling and top dressing could have pushed the yield to ~80 lbs, a 33% increase.

You can follow up with questions about specific parts, and Cursor pulls the relevant passage with timestamps so you can jump to that section in the video.

Wrapping Up

Once AgentPatch is connected to Cursor, you get the YouTube Transcript tool plus the full marketplace: web search, email, image generation, Google Maps, and more. One configuration, persistent access. Visit agentpatch.ai to see what’s available.

← Back to blog