{
  "markdown": "# Browserless Extension for Gemini CLI\n\nA Gemini CLI extension that connects Gemini to [Browserless.io](https://browserless.io) through the hosted **Browserless MCP server**. It gives Gemini an interactive web agent plus scraping, web search, crawling, site mapping, custom Puppeteer, Lighthouse audits, and export — all from natural language, with no local browser infrastructure to manage.\n\n## Installation\n\nClone the repo and link the extension:\n\n```bash\ngit clone https://github.com/browserless/gemini-extension.git\ncd gemini-extension\ngemini extensions link .\n```\n\n## Setup\n\n### 1. Get a Browserless API token\n\nSign up for free at [browserless.io](https://www.browserless.io) and grab your API token.\n\n### 2. Configure the token\n\nWhen you enable the extension, Gemini CLI prompts for the **API Token** setting and stores it as `BROWSERLESS_TOKEN`. Alternatively, set it in your shell before launching:\n\n```bash\nexport BROWSERLESS_TOKEN=your-token-here\n```\n\nThe token is sent as a `Bearer` header to `https://mcp.browserless.io/mcp`. That's the whole setup — the browser tools are served over MCP and appear automatically.\n\n### 3. Start using it\n\nJust ask. Gemini picks the right tool:\n\n```\nGo to news.ycombinator.com, open the top story, and summarize the linked article.\nTake a screenshot of https://example.com.\nFind recent AI news from the last week.\nCrawl https://browserless.io and list every page.\n```\n\n## Tools\n\nThese are exposed by the Browserless MCP server (the live set is loaded from the server at runtime):\n\n| Tool | Description |\n|------|-------------|\n| `browserless_agent` | Interactive web agent — drives a persistent browser session through a multi-step ReAct loop (navigate, click, type, snapshot). Handles multi-tab workflows, captcha solving, residential proxies, and file upload/download. |\n| `browserless_skill` | On-demand recipes for tricky page mechanics (cookie banners, modals, shadow DOM, dynamic content). Companion to the agent. |\n| `browserless_smartscraper` | Scrape a page with cascading strategies (HTTP fetch → proxy → headless browser → captcha solving). Returns markdown, HTML, or links. |\n| `browserless_search` | Structured web, news, and image search. |\n| `browserless_crawl` | Follow links from a seed URL to a configurable depth, scraping each page. |\n| `browserless_map` | Discover every URL on a site (sitemaps, pages, subdomains). |\n| `browserless_function` | Run custom Puppeteer JavaScript in a cloud browser. |\n| `browserless_performance` | Lighthouse audit (performance, accessibility, best practices, SEO). |\n| `browserless_export` | Export a page in its native format, optionally bundling assets into a ZIP. |\n\n## Extension Structure\n\n```\ngemini-extension/\n  gemini-extension.json     # Manifest — MCP server pointer + token setting\n  GEMINI.md                 # Context loaded at session start\n```\n\n## License\n\nSSPL-1.0\n",
  "bytes": 2878,
  "sha": "336b34ec40b8f7cebae1a0b97be9c948f0b60b3706fd1ae1954e952763bc5fbb",
  "repo_slug": "browserless/gemini-extension",
  "fonte": "repo",
  "truncated": false,
  "api": "https://agentalog.com/api/listings/plg_browserless_gemini_extension_e5249ff9/readme"
}