Skip to the collection

agent-native · Browser & Web Execution

Skyvern

Listed on The Agent-Native Index

AI agents to automate workflows on any website

What it does

Skyvern runs vision-driven browser automation: an agent describes a goal in natural language; Skyvern drives a real browser, plans actions, and completes multi-step workflows (forms, logins, downloads, structured extraction) without brittle XPath-only scripts. It supports CAPTCHA handling, 2FA/TOTP flows, geo-targeted proxies, persistent browser sessions, JSON-schema extraction, webhooks, and recordings for audit. Skyvern Cloud offers a production API; the core stack is open source for self-hosting.


Agent-facing interfaces

How agents onboard

Interaction pattern: SDK / REST + optional MCP

curl -X POST https://api.skyvern.com/v1/run/tasks \
  -H "x-api-key: $SKYVERN_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"prompt": "Find the top 3 posts on Hacker News", "url": "https://news.ycombinator.com"}'

See the full onboarding section in the dossier.

Protocol surface

Interface Detail
REST https://api.skyvern.com — OpenAPI-linked from docs
Python SDK https://www.skyvern.com/docs/sdk-reference/overview
TypeScript SDK https://www.skyvern.com/docs/ts-sdk-reference/overview
Self-hosted Docker / K8s — own LLM keys