Introducing /interact. Scrape any page, then let your agent take over to click, type, and extract data for you. Try it now →

How to summarize a YouTube video with Claude

Claude cannot visit URLs on its own. If you paste a YouTube link, it can only guess from its training data — it cannot pull the current transcript. The Firecrawl MCP server bridges this gap by giving Claude live web scraping through the Model Context Protocol.

Add the server to your Claude Desktop config:

{
  "mcpServers": {
    "firecrawl": {
      "command": "npx",
      "args": ["-y", "firecrawl-mcp"],
      "env": {
        "FIRECRAWL_API_KEY": "fc-YOUR_API_KEY"
      }
    }
  }
}

After restarting Claude, paste a YouTube URL and ask for a summary. Claude calls Firecrawl's scrape tool behind the scenes, receives the transcript as clean markdown, and returns a structured summary. You can follow up with requests to extract action items, list timestamps for specific topics, or translate the summary — Firecrawl delivers the full transcript text, so Claude has all the context it needs.

Firecrawl handles JavaScript rendering and page structure changes automatically, so the integration keeps working even when YouTube updates its frontend. Get a free API key at firecrawl.dev to start.

Last updated: Apr 06, 2026
FOOTER
The easiest way to extract
data from the web
Backed by
Y Combinator
LinkedinGithubYouTube
SOC II · Type 2
AICPA
SOC 2
X (Twitter)
Discord