Skyvern

Skyvern is an open-source browser agent that uses vision models to run workflows on websites without fixed selectors.

Skyvern automates browser workflows with large language models and computer vision instead of brittle XPath or DOM scripts. A swarm of agents reads the page, plans steps and acts through Playwright, so the same workflow can run on sites the system has not seen before. It ships a Playwright-compatible SDK and a no-code workflow builder. Skyvern Cloud is the managed service, with parallel runs, anti-bot measures, proxies and CAPTCHA solvers. The project can also be run locally with a UI and server. A technical report is published with the repository.

Skyvern at a glance

Best for: Software engineer, Founder
Runs on: Web, open source, self-hosted
Pricing: Free and open source to self-host; Skyvern Cloud is a paid managed service
Released: February 2024
Category: AI agent tool
Website: github.com/Skyvern-AI/skyvern

Key features

Vision LLM agents on top of Playwright
No-code workflow builder and a Playwright-compatible SDK
Works on sites without custom selectors
Skyvern Cloud with proxies and CAPTCHA solvers
Local UI and server install
Open-source core on GitHub

Use it for

Fill the same form on many vendor portals
Replace a brittle Selenium script after a redesign
Have non-engineers record a browser workflow

Alternatives to Skyvern

WebBrain — Runs locally or in the cloud with your choice of LLM and zero telemetry.
BrowserOS — Runs entirely locally and lets AI agents use real user logins to complete tasks.
Webwright — Enables code-based web automation with reusable and rerunnable scripts.
Nanobrowser — Run multi-agent workflows using your own LLM API key directly in your browser.
Steel — Run scalable cloud browser sessions with a simple developer API.
Stagehand — It combines familiar APIs with self-healing actions and token efficiency for agents.
Dev Browser — Persistent browser tabs keep state across scripts so agents do not restart.
Cloudbrowser — It gives every AI agent an isolated, persistent Chrome browser with saved login states.
Lightpanda — Lower memory footprint and faster execution than Chrome for automated workflows.
BrowserAct Skills — Builds tested web scrapers from single text prompts with managed infrastructure.
AI Browser Workflows — Renders alternative website layouts with reusable data.
browser-control — Shell-native design with stable element refs and no LLM or framework requirement.

All Skyvern alternatives