Persistent sessions for remote AI coding agents
How persistent terminal sessions let remote AI coding agents keep long-running work available between client connections.
On this page
SHORT ANSWER
How do persistent sessions work for remote AI coding agents?
A Persistent Agent Session keeps terminal work attached to the remote cloud computer instead of a single browser or SSH connection. You can disconnect one client and later reconnect to the same session from another device while the remote cloud computer remains online.
This guide explains how persistent/remote AI agent sessions work. For the product capability that keeps your TermSquad workflows available between connections, see Persistent Agent Sessions.
Key takeaways
- A Web Terminal or SSH connection is an access method, not persistence by itself.
- Persistent Agent Sessions keep long-lived terminal work on the remote cloud computer between client connections.
- The remote cloud computer must remain online for the session and its processes to remain available.
- Reconnecting from another device returns you to the remote environment rather than moving the workload to that device.
Access and persistence are different layers
The client is how you reach the machine. A browser terminal, an SSH client or a compatible Remote SSH desktop workflow can all be clients. The persistent session is a separate layer that stays on the remote machine.
A typical remote workflow
- Connect to the remote cloud computer.
- Open or create a Persistent Agent Session.
- Start the coding agent inside that session.
- Disconnect the client when you need to leave.
- Reconnect later and attach to the existing session.
Why this matters for agent work
Coding agents often perform tasks that outlast a single interactive check-in. Separating the persistent session from the client means the workflow is tied to the remote cloud computer rather than to whichever laptop, browser or SSH client you used to start it.