Skyvern

Builds Skyvern’s browser workflow platform and coding-agent interfaces, with cloud and self-hosted execution options.

About Skyvern

Skyvern-AI maintains the Skyvern browser-workflow project. Its source and documentation cover both direct browser control and workflows, with agent-facing MCP and CLI entry points rather than a separate Skill-only product.

Skyvern combines direct browser operations with reusable browser workflows. A one-off lookup can start from a URL and requested fields; a recurring process benefits from an explicit workflow with run inputs and a reviewable result. The hosted service and self-hosted stack differ in authentication and operating responsibility.

Cloud service access and self-hosted software share a project identity but have different operating requirements. Official Skills help use the selected interface; the project guide explains their setup and execution requirements.

Software from Skyvern

Distinguish a one-off browser action from a workflow that will run repeatedly. That decision determines whether the first input is a page or an existing workflow and its run data.

Skyvern
Use Skyvern’s MCP or API for browser operations and multi-step workflows. Choose cloud or self-hosted execution, supply the required credentials, and review the business result rather than only the run status.
- Deployment, Browser & Desktop
- Browser Automation
- Browser Automation
- MCP

Alternatives to Skyvern Workflows

Compare Browser Use for model-driven browser tasks, Browserbase for SDK automation and hosted sessions, Bright Data for managed web data services, UncleCode for content crawling and extraction. Choose by the browser environment and task result you need.

Frequently Asked Questions

Who publishes the Skyvern platform?

The linked upstream repository belongs to Skyvern-AI. Use its official project and documentation links to identify current releases and interfaces.

Is cloud access the only way to use it?

No. Official documentation describes self-hosted and cloud paths. Their authentication, model, browser, and maintenance responsibilities differ.

Are the CLI Skills upstream-maintained?

The skyvern instructions are in the official repository. They accompany the runtime and do not replace backend setup or a service account.

What is the difference between authoring and running?

Authoring changes a workflow definition; operating executes it with inputs. Treat those permissions and verification requirements separately.

What terms govern the source?

The repository is AGPL-3.0. Hosted subscriptions and model charges are separate from source licensing and require their own review.

Where should session errors be investigated?

Use the installed MCP/CLI release’s documentation, session identifiers, and backend logs. Do not mix an older integration guide with current session-creation requirements.

Review a One-Record Workflow

Run one existing workflow or one direct browser task first. This reveals missing inputs without silently turning a test into workflow-authoring work.

Open Skyvern guide