01Integrates with Model Context Protocol (MCP)
02Automates browser interactions (navigation, form filling, etc.)
03Utilizes visual understanding through screenshot analysis
04Persists browser sessions between tasks
05Supports multiple LLM providers (OpenAI, Anthropic, Azure, etc.)