Browserless

Author: browserless Type: tool Repository: https://github.com/browserless/dify-plugin

Headless browser automation for AI agents, powered by Browserless. This plugin proxies the official Browserless MCP server (https://mcp.browserless.io/mcp) and exposes its tools as native Dify tools.

What you can do

• Scrape any web page and get markdown, HTML, screenshots, PDFs, or links
• Search the web and optionally scrape every result
• Map every URL on a website (sitemap + link discovery)
• Crawl a site and scrape every page
• Export a URL in its native content type
• Run custom Puppeteer code server-side and get the result back
• Download files triggered by Puppeteer code
• Audit performance with Lighthouse
• Drive a stateful agent (goto, click, type, snapshot, etc.) — full ReAct browser loop

Setup

1. Get a Browserless API token by signing up at browserless.io/signup/email.
2. In Dify: Tools → Browserless → Authorize, paste the token, save.
3. Add any Browserless tool to an Agent or Workflow node.

Tools

Tool	Use case
browserless_smartscraper	Extract page content as markdown / HTML / screenshot / PDF / links
browserless_search	SearXNG-backed web/news/image search, optional inline scrape
browserless_map	Discover every URL on a site
browserless_crawl	Crawl + scrape a site up to a depth
browserless_export	Fetch a URL in its native content type (HTML, PDF, image…)
browserless_function	Run arbitrary Puppeteer JS, return a typed payload
browserless_download	Run Puppeteer JS, return a file Chrome downloads
browserless_performance	Run a Lighthouse audit
browserless_agent	Persistent multi-turn agent session (goto, click, type, snapshot, …)

Privacy

See PRIVACY.md. Browserless privacy policy: https://www.browserless.io/privacy.

Support

• Issues: https://github.com/browserless/dify-plugin/issues
• Email: support@browserless.io