Skyvern
Browser workflow automation driven by LLMs and computer vision
What it is
An agent that completes tasks on websites by reading the rendered page with vision models instead of relying on fixed selectors, so flows survive layout changes. It offers a Playwright-compatible SDK and a no-code workflow builder for chaining steps like logins, form filling and downloads. For teams automating portals that have no API.
Good for
- Fill forms on sites without an API
- Download invoices from vendor portals
- Chain multi-site browser tasks into a workflow
Install or run
pip install "skyvern[all]"Command as given in the project README. Read the repository before you run anything on your machine.
What you need
Python, a browser runtime and a vision-capable LLM API key
Sources
- https://github.com/Skyvern-AI/skyvern (project README, read 11 Nov 2026)
- https://api.github.com/repos/Skyvern-AI/skyvern (GitHub API, read 13 Nov 2026)
The summary is ours, written from the README. Spotted an error? See how we work.