A natural language browser automation skill that enables AI agents to navigate web pages, extract data, and interact with web elements via the CLI.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install stagehand-browser-cli
2. Manual Installation
Copy the skill folder to one of these locations
~/.openclaw/skills/ <project>/skills/ Priority: Workspace > Local > Bundled
3. Prompt Installation
Copy this prompt to OpenClaw to install it automatically.
Help me install stagehand-browser-cli using Clawhub. If Clawhub is not installed, install it first (npm i -g clawhub).
Prefer to download?
Get the raw skill files in a ZIP archive.
What is Browser Automation?
The Browser Automation skill provides a powerful interface for controlling web browsers using natural language instructions. By integrating the Stagehand CLI, it allows users and AI agents to execute complex web tasks without writing brittle selector-based scripts. This skill is a core component of the Openclaw Skills ecosystem, offering a bridge between high-level intent and low-level browser actions.
One of the standout features of this skill is its flexible environment management. It can automatically detect whether to run sessions locally using a standard Chrome installation or remotely via Browserbase. This makes it ideal for everything from local development and testing to production-scale web scraping and automation in environments where stealth and proxy support are required.
Browser Automation Use Cases
- Automating data extraction from complex websites using natural language schemas.
- Navigating through multi-step web forms and applications for automated testing.
- Capturing periodic screenshots of web interfaces for visual monitoring.
- Using natural language to discover and interact with dynamic page elements via the observe command.
- Building agentic workflows that require real-time web research and information retrieval as part of the Openclaw Skills framework.
How Browser Automation Works
- The skill initializes by checking the .env file for Browserbase credentials to determine the execution environment.
- It triggers a browser instance (local or remote) based on the available configuration.
- The navigate command is used to point the browser to the target URL.
- Commands like act or extract are processed through Stagehand to translate natural language into DOM interactions.
- Screenshots or observation logs are generated to verify the state of the browser.
- The session is concluded with the close command to release system or cloud resources.
Browser Automation Setup
To integrate this capability into your environment, follow these steps to set up the browser tool within the Openclaw Skills directory:
npm install
npm link
To enable remote execution with stealth features, add your credentials to the .env file:
BROWSERBASE_API_KEY=your_keyBROWSERBASE_PROJECT_ID=your_project_id
Browser Automation Data Schema & Taxonomy
The skill manages browser states and provides structured output for data extraction tasks. Below is the metadata and command structure:
| Type | Command | Description |
|---|---|---|
| Navigation | navigate <url> |
Loads the specified web page. |
| Interaction | act <instruction> |
Executes a natural language action like 'click' or 'type'. |
| Extraction | extract <instruction> |
Returns structured data from the page based on the query. |
| Inspection | observe <query> |
Returns a list of interactive elements matching the query. |
| Utility | screenshot |
Captures a PNG of the current browser viewport. |
Browser Automation Advanced Features
- Automatic environment switching between local Chrome and remote Browserbase sessions.
- Support for stealth mode and CAPTCHA solving when running in remote production environments.
- Natural language element discovery that eliminates the need for maintaining fragile CSS selectors or XPaths.
- High-performance data extraction that can be enforced with custom JSON schemas.
- Seamless interoperability with other Openclaw Skills for complex, multi-agent automation pipelines.
SKILL.md
Loading
Related Openclaw Skills
Browserbase Functions: Serverless Browser Automation
A comprehensive skill for deploying, managing, and invoking serverless browser automation using the Browserbase CLI and SDK.
Browser Automation Skill
A natural language interface for automating browser actions, web navigation, and data extraction via CLI.
Browser Automation
An AI-powered browser automation tool that executes natural language commands to navigate, interact with, and extract data from any website.
Browser Automation
A powerful CLI skill for automating web browser interactions and data extraction using natural language commands.
Capability Awareness System
An automated discovery system that allows agents to identify and load custom skill documentation on demand.
Amazon Orders Skill
A robust tool for downloading and querying Amazon order history through an unofficial Python API and CLI interface.
Featured*
