A stable, native WKWebView browser for OpenClaw agents that supports authenticated web sessions and tab management without requiring search API keys.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install openclaw-browser-2
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 openclaw-browser-2 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 OpenClaw Native Browser?
The OpenClaw Native Browser is a high-performance WKWebView implementation designed to give AI agents a real presence on the web. Unlike basic headless scrapers, this is one of the Openclaw Skills that provides a visible window with full tab management and a native URL bar. It allows agents to interact with modern, complex web applications like Perplexity, Grok, and ChatGPT just like a human would, bypassing the limitations of standard search APIs.
By incorporating this into your collection of Openclaw Skills, you eliminate the need for paid API keys for web searching. The browser maintains persistent cookies and sessions, enabling your agent to stay logged into services across multiple tasks. Its two-window architecture ensures that the agent can manage navigation via a toolbar while interacting with specific content windows through a clean IPC interface.
OpenClaw Native Browser Use Cases
- Navigating and interacting with authenticated web applications like Claude or ChatGPT
- Performing web searches and research without relying on external API keys
- Automating complex multi-step workflows across different browser tabs
- Capturing screenshots and DOM snapshots for data extraction and analysis
- Handling websites that require real user interaction and JavaScript execution
How OpenClaw Native Browser Works
- The OpenClawBrowserSkill is initialized as a singleton, making it a powerful addition to your Openclaw Skills.
- A dedicated child process is spawned to manage the pywebview main thread, ensuring the browser doesn't block the agent's logic.
- A frameless toolbar window is created to provide a visual UI for tab and URL management.
- For every new tab, a native WKWebView content window is instantiated.
- Commands are passed via JSON-based IPC over stdin/stdout to execute JavaScript, click elements, or type text.
- The browser returns data, titles, or snapshots back to the agent for processing.
OpenClaw Native Browser Setup
To integrate this tool with your Openclaw Skills, clone the repository and install the package using pip:
git clone https://github.com/yungookim/openclaw-browser.git ~/clawd/openclaw-browser
cd ~/clawd/openclaw-browser
pip install -e .
Verify the installation by running the following command in your terminal:
python -c "import sys; sys.path.insert(0, '/Users/$USER/clawd/openclaw-browser'); from src import OpenClawBrowserSkill, __version__; print(f'openclaw-browser v{__version__} ready')"
OpenClaw Native Browser Data Schema & Taxonomy
The skill utilizes a structured API to manage navigation and interaction. Below is a summary of the data and methods available within this addition to your Openclaw Skills:
| Feature | Method | Purpose |
|---|---|---|
| Navigation | load(url) | Loads a URL in the active tab |
| State | snapshot() | Returns HTML content and metadata |
| Interaction | click(selector) | Performs a native click on a CSS selector |
| Input | type_text(selector, text) | Types text into specific input fields |
| Tabs | new_tab(url) | Opens and tracks a new browser instance |
| Auth | login_claude() | Executes a specialized login helper flow |
OpenClaw Native Browser Advanced Features
- Built-in login helpers for major AI platforms including Perplexity, Grok, Claude, and ChatGPT
- Multi-tab orchestration with independent switching and closing capabilities
- Persistent session management via native macOS cookie storage for long-running tasks
- Frameless UI architecture for a clean, non-intrusive developer experience during agent execution
- Subprocess isolation ensures the browser runs in a child process, preventing crashes from affecting the main agent thread
SKILL.md
Loading
Related Openclaw Skills
Yumstock: US Stock Analysis Agent
A sophisticated US stock analysis agent that uses a three-pillar weighted scoring system and macro-environment gating to provide data-driven Buy, Hold, or Sell verdicts.
Task Planning
A professional agile framework for breaking down complex software features into manageable user stories and actionable development tasks.
Remembering Conversations
A sophisticated context-retrieval skill that empowers AI agents to search, synthesize, and apply insights from past conversations to solve current problems.
Market Research Reports Agent
A sophisticated AI skill designed to generate 50+ page, consulting-grade market research reports with professional LaTeX styling and extensive visual analytics.
Openclaw X Article Cover Generator
An AI-powered generator designed to create high-quality, brand-consistent X (Twitter) article covers featuring the OpenClaw lobster logo.
Brainstorming
A mandatory pre-implementation skill for exploring user intent and creating technical designs through collaborative dialogue.
Featured*
