Skyvern

Skyvern is an open-source AI agent platform built by Ikonomos Inc. that automates browser-based workflows on any website, without relying on brittle CSS or XPath selectors. Instead, it combines computer vision with large language models to understand a page visually, similar to how a person would, so it can adapt automatically when a site's layout changes. Users define workflows through a no-code or low-code builder, or trigger them programmatically via Python and TypeScript SDKs and a REST API, making it possible to run thousands of tasks in parallel. The platform is aimed at teams that need to replace manual, repetitive web interactions such as downloading invoices, extracting structured data, filling out forms, or completing login flows that involve CAPTCHA solving and two-factor authentication. It supports proxy networks with geographic targeting for region-specific automation and produces JSON or CSV data output alongside document processing and OCR capabilities. Every action the agent takes is logged with a real-time, human-readable summary, giving Skyvern's "explainable AI" a degree of auditability that many black-box automation tools lack. Skyvern integrates with Zapier, Make, Clay, n8n, and Workato, and can be self-hosted via Docker for organizations with stricter data requirements. It offers a free tier alongside enterprise plans available on request. One limitation: because it depends on live visual interpretation of pages, complex or heavily obfuscated sites may still require workflow tuning.