openclaw-browsy

Zero-render browser plugin for OpenClaw. Integrates browsy as a fast, lightweight alternative to Playwright/CDP for AI agent browsing tasks.

Why browsy?

OpenClaw's built-in browser uses Playwright + CDP: ~300MB RAM, 2-5s per page. browsy handles 70%+ of agent browsing tasks (forms, logins, search, data extraction) at 10x speed and 60x less memory by parsing HTML into a Spatial DOM without rendering pixels.

Install

npm install openclaw-browsy

Requires the browsy CLI binary in your PATH. Install from the browsy releases.

Configure

In your OpenClaw config:

{
  "plugins": {
    "openclaw-browsy": {
      "port": 3847,
      "autoStart": true,
      "allowPrivateNetwork": false,
      "preferBrowsy": true,
      "serverTimeout": 10000
    }
  }
}

| Option | Default | Description | |--------|---------|-------------| | port | 3847 | Port for the browsy REST server | | autoStart | true | Start browsy serve automatically on plugin init | | allowPrivateNetwork | false | Allow fetching private/internal network URLs | | preferBrowsy | true | Intercept built-in browser tool calls and redirect through browsy | | serverTimeout | 10000 | Timeout (ms) waiting for server startup |

Usage

As an OpenClaw plugin

// openclaw.config.ts
import { register } from "openclaw-browsy";
export default { register };

Once registered, every agent gets 13 browsy tools automatically:

| Tool | Description | |------|-------------| | browsy_browse | Navigate to a URL | | browsy_click | Click an element by ID | | browsy_type_text | Type text into an input | | browsy_check | Check a checkbox/radio | | browsy_uncheck | Uncheck a checkbox/radio | | browsy_select | Select a dropdown option | | browsy_search | Search the web | | browsy_login | Log in with credentials | | browsy_enter_code | Enter a 2FA/verification code | | browsy_find | Find elements by text or ARIA role | | browsy_page_info | Get page metadata and suggested actions | | browsy_tables | Extract structured table data | | browsy_back | Go back in browsing history |

Standalone

import { BrowsyContext } from "openclaw-browsy";

const ctx = new BrowsyContext({ port: 3847, autoStart: false });
const result = await ctx.executeToolCall("browse", { url: "https://example.com" });
console.log(result);

Skills

Three runtime skills are included:

• browse-and-extract — Navigate + extract with auto-handling of login walls and cookie consent
• web-research — Search + read multiple pages + compile summary
• form-filler — Detect form fields + fill + submit using page intelligence

License

MIT