Launch HN: Skyvern (YC S23) – open-source AI agent for browser automations

Skyvern is an open-source AI agent platform that automates browser-based workflows using LLMs, allowing users to define goal-based prompts to complete complex tasks across websites without brittle code-based solutions. The platform features a React UI for real-time monitoring, workflow chaining, authenticated sessions with 2FA support, and cached workflows for reusable interactions, with token costs reduced 80% using GPT-4O.

Key Points

• •Skyvern is an open-source AI agent that automates browser-based workflows using LLMs with goal-based prompts instead of brittle code-based scripts
• •Solves the problem of expensive manual operations teams or developer-heavy UI automation tools (like UIPath/Selenium) that break when websites change
• •Supports complex multi-step workflows across diverse websites: insurance quotes, job applications, permit filing, invoice fetching, e-commerce purchasing, and form submissions
• •Features real-time React dashboard visualization, livestreaming browser instances, authenticated session management (Bitwarden integration), and 2FA support (Email/Phone/QR-code)
• •Implements workflow chaining to handle multi-step processes and cached workflows that memorize previous interactions for cost-efficient re-use
• •Processes complex HTML elements (SVG identification/summarization) and handles dynamic website interactions like autocomplete iteration for accurate data entry
• •Reduced token costs 80% by switching from GPT-4V ($15/1M tokens) to GPT-4O ($2.50/1M tokens), with $5 free credits for new users to test
• •Open-source repository available on GitHub with cloud version at app.skyvern.com for immediate deployment without local infrastructure

Workflow Diagram

Concepts