A secure web sandbox for AI agents to browse tool
Browser Sandbox provides a secure, isolated browser environment that lets AI agents interact with live websites while keeping your systems safe. It enables real-time data extraction and testing, delivering LLm-ready data (Markdown, JSON, screenshots) with support for Python, Node.js, and CLI workflows, so your AI apps access clean web data efficiently.
Browser sandboxes for your agents to interact with the web.
- Install the CLI: npx -y firecrawl-cli@latest init --all --browser
- Initialize your agent to use the browser sandbox and start scraping, searching, crawling, and mapping
- Example (Python):
```python
from firecrawl import Firecrawl
app = Firecrawl(api_key='YOUR_API_KEY')
app.scrape('example.com')
```
Secure, isolated sessions prevent cross-agent data leaks
Open-source and transparent components
Supports Markdown, JSON, and screenshots output
Easy CLI and SDK integration for developers