Skyvern

Skyvern uses computer vision and large language models to drive a browser the way a human would. Instead of relying on CSS selectors or XPath expressions that break when a website changes, it visually understands page elements and adapts to layout updates on its own. You can build workflows by describing them in plain English, recording your actions, uploading SOP documents, or using the visual drag-and-drop editor.

The tool handles real-world complications that trip up traditional automation scripts. It solves CAPTCHAs natively, manages two-factor authentication flows, supports proxy networks for geo-targeted browsing, and handles dynamic form fields automatically. Workflows can be triggered through the visual builder, REST API, webhooks, or the Python and TypeScript SDKs.

Skyvern runs in the cloud or self-hosted via Docker containers. The open-source version is available on GitHub under AGPL-3.0 with over 20,000 stars. You can bring your own LLM provider, including OpenAI, Anthropic, Google Gemini, or local models through Ollama.