Browser & Shell Access
How the built-in browser and terminal work inside every agent.
Every agent on x3s.host includes two built-in access tools alongside the runtime itself: a headless browser and an in-dashboard terminal. Both are ready immediately after deployment — no setup needed.
Built-in browser
Each agent runs with a browser sidecar. Your runtime can use it to:
- Load and interact with web pages
- Automate browser-based workflows
- Run tasks that require a real browser (screenshots, form submission, scraping)
The browser is available to the runtime automatically. You don't need to configure it — if your runtime supports browser tasks, it will use the sidecar.
You can also open the browser directly in a new tab from the agent panel using the Open Browser link. This gives you a live view of what the browser is currently showing.
In-dashboard Terminal
The Terminal tab gives you a real shell session inside your agent container. It's available from the agent detail panel and runs directly in your browser.
Use it to:
- Run commands to check your agent's state
- Read log files or check environment variables
- Test things out without leaving the dashboard
The terminal is connected to the same container your runtime is running in, so you have full access to the runtime's environment.
Tailscale SSH
For a persistent shell session from your own machine, connect Tailscale and use SSH:
ssh <username>@<tailscale-ip>This gives you the same access as the in-dashboard terminal, but from your local terminal with full support for your preferred tools, multiplexers, and key-based auth. See Networking & SSH for setup details.
If browser-based tasks fail
- Check your agent status — it should be
ONLINE - Open Logs and look for errors at startup
- Try the operation again — browser startup occasionally lags a few seconds behind the runtime
- If it keeps failing, use Restart from the agent panel and try again once the agent is back online