Devin Desktop now keeps coding agents alive over SSH — agentic development is quietly becoming a remote-server habit
Cognition's Devin Desktop (formerly Windsurf) shipped SSH connectivity with automatic session recovery on 10 September 2026, followed by reliability fixes on 15 September — letting developers leave AI coding agents running on a remote server and reconnect from any machine without losing the session.
17 September 2026
Cognition’s Devin Desktop — the product formerly known as Windsurf — added SSH connectivity to remote agents with automatic session recovery on 10 September 2026, then shipped reliability improvements to that same reconnection flow just five days later. The practical effect: an AI coding agent can run persistently on a cloud server or a beefier remote machine, and a developer can close their laptop, open it later, or switch devices entirely, and pick the session back up where it left off.
A small feature that points at a bigger shift
This isn’t an isolated move. Cursor has been building out cloud agents for parallel work away from a local machine, and Claude Cowork has been expanding into browser and mobile contexts. Put together, the direction is clear: the coding agent itself is decoupling from any one developer’s laptop and becoming a persistent process a team checks in on, closer to how a CI pipeline or a long-running background job behaves than how an IDE plugin behaves. For teams distributing work across time zones or wanting agents to keep working overnight, that’s a genuine capability upgrade — but it also means the agent now has standing access to a remote environment between sessions, which is a new thing to secure and monitor, not just a new thing to use.
So what
If your development team — internal or outsourced — is running AI coding agents on persistent remote sessions, ask how those sessions are secured when nobody’s actively watching them, and who’s accountable for what an agent does between check-ins. That accountability question is exactly why we scope AI-assisted work with clear human review points built in, whatever infrastructure the agents run on — read more on our AI-assisted development page, or get in touch to talk through how we’d structure it for your team.