Skip to main content
Antigravity IDE is Google’s AI-powered code editor. It combines a fully featured editor with an agent that operates across your editor, terminal, and browser, handling tab autocompletion, natural-language code commands, and autonomous task execution. With ZenRows connected, your coding agent can fetch live documentation, pull real API specs, and scrape technical references without getting blocked.

​
Prerequisites

​
Setup

1

Step 1: Open the configuration file

Antigravity IDE shares its configuration file with Antigravity 2.0. Open the MCP configuration file:
~/.gemini/config/mcp_config.json
If the file does not exist yet, create it.
2

Step 2: Add the ZenRows MCP server

Add the following configuration, replacing YOUR_ZENROWS_API_KEY with your actual API key:
JSON
{
  "mcpServers": {
    "zenrows": {
      "command": "npx",
      "args": ["-y", "@zenrows/mcp"],
      "env": {
        "ZENROWS_API_KEY": "YOUR_ZENROWS_API_KEY"
      }
    }
  }
}
3

Step 3: Restart Antigravity IDE

Save the file and restart Antigravity IDE to load the new configuration. ZenRows tools will be available to the agent immediately.
For best results with MCP tool calls, use Claude Opus or Claude Sonnet as your model via the model selector in the IDE.

​
Fetch live data in your editor

With ZenRows connected, ask the agent to fetch and work with live web content directly in the editor. To scrape a page and summarize its contents: 
Scrape https://www.scrapingcourse.com/antibot-challenge and tell me what the page is all about.
For a more focused use case: 
Use ZenRows to scrape https://github.com/vercel/next.js/issues and find open issues related to hydration errors. Summarize the most common causes and suggested fixes.

​
Using autonomous agents with ZenRows

Antigravity IDE includes an agent that handles longer, multi-step tasks across your editor, terminal, and browser. For scheduled tasks, parallel research workflows, or autonomous data extraction, see the ZenRows + Antigravity 2.0 integration guide for the full agent setup.

​
Troubleshooting

​
ZenRows tools not appearing after saving the config

  1. Confirm the file is saved to ~/.gemini/config/mcp_config.json.
  2. Restart Antigravity IDE to force a config reload.
  3. Check that the JSON is valid, with no trailing commas or comments.

​
Agent not using ZenRows tools

Antigravity IDE may use its built-in browser tool instead of ZenRows. Instruct the agent explicitly: 
Use ZenRows to scrape https://www.scrapingcourse.com/antibot-challenge with JavaScript rendering.

​
JSON syntax error on startup

The configuration file must be valid JSON. Run it through a validator if Antigravity IDE fails to load the server.

​
FAQ (Frequently asked questions)

Yes. Both tools can coexist in the same session. Antigravity IDE picks its built-in browser tool by default for general browsing tasks, so when you need ZenRows specifically — for anti-bot bypass, JavaScript rendering, or residential proxies — instruct the agent explicitly to use ZenRows. See the troubleshooting section for the exact prompt pattern.
Yes. Each request the agent makes through ZenRows counts against your plan credits. JavaScript rendering (js_render) and Premium Proxies consume more credits per request than a standard fetch. You can monitor your usage in real time from the ZenRows dashboard.
Yes. The mcp_config.json file path uses Unix-style notation in this guide, but on Windows the equivalent path is %USERPROFILE%\.gemini\config\mcp_config.json. The rest of the setup is identical.
Yes. Add the zenrows block inside the existing mcpServers object alongside any other servers already defined. Each server runs independently, so adding ZenRows does not affect your other MCP connections.
JSON
{
    "mcpServers": {
        "your-existing-server": { ... },
        "zenrows": {
            "command": "npx",
            "args": ["-y", "@zenrows/mcp"],
            "env": {
                "ZENROWS_API_KEY": "YOUR_ZENROWS_API_KEY"
            }
        }
    }
}