Repository navigation
Home
Dr K edited this page Nov 14, 2026
·
13 revisions
Drive a real Chrome or Firefox on Selenium Grid from an agent over MCP, or from anything else over plain HTTP. The browser stays open between calls, with its page, cookies and scroll position.
The README is the tour, and its quick start runs the server locally in two commands. These pages say when and how to use each part. The exact rules are in the spec.
| You want to | Read |
|---|---|
| Connect a client: Claude, VS Code, n8n, Cursor, curl | Installing |
| Run the server: stdio, HTTP, Docker, Kubernetes | Deployment |
| Set an option | Configuration |
| Name a workspace, or recover a dead browser | Workspaces |
| Make a token, use the admin UI, share a file | Administration |
| Look up one action's parameters | Actions |
| Save a sequence and replay it in one call | Flows |
| Type a password you are never shown | Secrets |
| Keep a download or a screenshot past the browser that made it | Files |
The browser lives on the Grid, and the workspace you name leads back to it, so the server can restart without anyone losing a tab.
- State carries over. Sign in once and every later call is signed in. A browser left on the wrong page makes your next selector fail for a reason that has nothing to do with the selector.
- The Grid runs a handful of browsers, and an abandoned one holds its slot until it is reaped. Close it when you are done.
The action pages are generated from the server's tool schemas, so they describe the version that shipped.
selenium-flow · MIT
Start here
Guides
Lifecycle
Going places
Doing things
Getting things out
Site data
Console and network