Skyvern

AI agents that automate browser workflows using LLMs and computer vision

4.5 (4)

Перевірено Daniel Nikulshyn·Оновлено травень 2026 р.

Open Source Workflow Automation Computer Vision LLM Self-Hosted API Developer Tools Browser Automation

Огляд

Skyvern is an open-source tool that uses large language models and computer vision to automate tasks in a web browser. Instead of relying on brittle XPath selectors or hand-coded scripts, it interprets pages visually and semantically, allowing it to adapt when site layouts change. It is aimed at developers and operations teams that need to automate repetitive web tasks such as form filling, data extraction, procurement workflows, and account management across many sites. Skyvern can be self-hosted or used via its managed cloud, and exposes an API for integrating browser automation into larger systems. Because it reasons about pages rather than memorizing structure, a single Skyvern workflow can often run across multiple similar websites without per-site customization.

Ключові функції

LLM and vision-based page understanding
Workflow builder for multi-step tasks
REST API and Python SDK
Handles CAPTCHAs and 2FA flows
Structured data extraction
Self-hosted or cloud deployment

Кейси використання

Automate Repetitive Form Filling

Use Skyvern to fill out web forms across many different sites without writing custom scripts for each, adapting automatically when page layouts change.

Structured Data Extraction at Scale

Extract structured data from web pages using visual and semantic understanding, useful for research, price monitoring, or aggregating information across sources.

Procurement and Account Management

Automate procurement workflows and routine account management tasks across vendor portals, including flows that involve CAPTCHAs or 2FA verification.

Embed Browser Automation via API

Integrate Skyvern's REST API or Python SDK into larger systems to add resilient browser automation as a backend service within existing applications.

Плюси і мінуси

Плюси

Adapts to website layout changes
Open-source with self-hosting option
Works across many sites without custom scripts
API-first for easy integration

Мінуси

Requires technical setup and LLM costs
Slower than scripted automation for simple tasks
Reliability varies on complex or dynamic sites

Відгуки

4.5

Середнє з 4 оцінок.

Увійди, щоб залишити відгук.

Pierre Dubois

Compared a few options

Evaluated this against two competitors. Where it wins: lLM and vision-based page understanding and open-source with self-hosting option. Where it lags: slower than scripted automation for simple tasks. On balance the feature set — especially handles CAPTCHAs and 2FA flows — justifies the 4 stars for our use case.

Marcus Bell

Years in this space

I've evaluated a lot of these over the years. What stands out here is workflow builder for multi-step tasks — handled better than most — and adapts to website layout changes. Reliability varies on complex or dynamic sites is my one real gripe. Worth the time if this is your use case.

Daniel Schmidt

Use it every day

Honestly didn't expect to like it this much. LLM and vision-based page understanding is exactly what I needed, and works across many sites without custom scripts. I do wish reliability varies on complex or dynamic sites, but I reach for it almost every day now and it just clicks.

Carlos Mendoza

Does the job

Pretty happy overall. Handles CAPTCHAs and 2FA flows just works and aPI-first for easy integration. Reliability varies on complex or dynamic sites can be annoying, but no dealbreakers — I'd recommend it to a friend without hesitating.