Skyvern

Open Source

Open-source AI agent that automates browser workflows using LLMs and computer vision instead of brittle scripts, handling CAPTCHAs and 2FA.

open sourcebrowser automation

Visit Skyvern →

Work at Skyvern? Manage this listing

Our take

An open-source browser-automation agent that uses vision and LLMs to survive UI changes that break traditional selector-based scripts.

Best for

Developers automating repetitive web workflows that break when sites change their UI.

Pros

Open-source (AGPL-3.0) with 20k+ GitHub stars
Uses computer vision + LLMs, not brittle CSS selectors
Survives UI changes without rewriting scripts
Handles CAPTCHAs and 2FA; cloud option at $0.05/step

Cons

Requires technical setup and oversight
AGPL licensing needs review for commercial use
Browser automation can be fragile on edge cases

How it compares

Versus traditional Selenium/Playwright scripts, Skyvern adapts to UI changes automatically; versus MultiOn, it is open-source and self-hostable.

Full review

Skyvern is an open-source AI agent that automates browser-based workflows using large language models and computer vision, replacing brittle, selector-based scripts that break on UI changes.

Because it reasons over the rendered page, it can adapt when sites change and can handle obstacles like CAPTCHAs and two-factor authentication; a paid cloud option adds infrastructure and anti-bot features at $0.05 per step.

It is a strong fit for engineering teams that want self-hostable, resilient automation, with AGPL licensing and setup overhead to weigh for commercial use.