skyvern

Skyvern is an AI-powered browser automation tool that leverages Vision LLMs and computer vision to interact with websites. Unlike traditional automation that relies on DOM parsing and XPaths, Skyvern can operate on unfamiliar websites and adapt to layout changes by reasoning through visual elements. Key capabilities include: - AI-Powered SDK: A Playwright extension adding natural language commands such as `page.act()`, `page.extract()`, and `page.validate()`, as well as AI-augmented versions of standard Playwright actions (click, fill, select, upload). - No-Code Workflow Builder: Allows users to chain multiple tasks together using blocks for data extraction, validation, loops, and HTTP requests. - Advanced Browser Control: Supports controlling local Chrome instances via remote debugging and tunneling local browsers to the Skyvern Cloud. - Data Extraction & Form Filling: Capable of extracting structured data based on a provided JSON schema and filling out complex web forms. - Authentication Support: Includes 2FA support (TOTP, Email, SMS) and integrations with password managers like Bitwarden. - Integrations: Supports Model Context Protocol (MCP) and connects with automation platforms such as Zapier, Make.com, and N8N. Skyvern can be deployed via a managed cloud service, local pip installation, or Docker Compose. It supports a wide range of LLM providers including OpenAI, Anthropic, Gemini, and locally hosted models via Ollama.