Communicating with Your Agent

There are four ways to interact with your agent. Use whichever fits the task.

Public URL — day-to-day use

Every agent gets a public https:// URL. This is the main way to use your runtime:

• Open it in a browser and sign in with your runtime username and password
• Use it as an API endpoint for integrations, passing the runtime token in your request headers

The Public URL is always available as long as your agent is ONLINE.

In-dashboard Terminal — quick diagnostics

The Terminal tab in your agent panel opens a live shell inside your container. Use it for:

• Checking that environment variables and secrets loaded correctly
• Running a quick command to validate runtime state
• Investigating an issue without leaving the dashboard

No setup required — it works immediately from the browser.

Tailscale SSH — private machine-to-agent access

When Tailscale is connected, you can SSH into your agent directly from your own machine:

ssh <username>@<tailscale-ip>

Use the username and password from Access & Credentials. The Tailscale IP (shown as tailscaleIp in the dashboard) is typically in the 100.x.x.x range.

This is the right path when you want a persistent terminal session or need to transfer files.

Logs — understanding what happened

The Logs tab shows streaming output from your agent. Check it when:

• Startup takes longer than expected
• The runtime behaves unexpectedly after adding secrets
• You want to confirm a request reached the agent

When something isn't working

1. Confirm agent status is ONLINE (not DEGRADED or STOPPED)
2. Re-check URL, username, password, and token from Access & Credentials
3. Open Logs and look for errors in the last few lines
4. Open Terminal and verify the runtime process is running
5. If using Tailscale, confirm the sidecar is connected and tailscaleIp is visible
6. If nothing helps, try Restart from the agent detail panel