Devin Cloud Comes to the Terminal: Command-Line Access to Cloud AI Coding Sessions

AI coding assistant Devin has shipped a developer workflow update that brings Devin Cloud capabilities directly into the terminal. Developers can now create, steer, resume, and watch cloud sessions from the command line — no more switching back and forth between the browser and their IDE.

Direct Terminal Access to Cloud Sessions

According to the official announcement, users can now manage Devin Cloud sessions right inside their preferred terminal. The workflow is straightforward: use the devin --cloud command or type /cloud in the interactive interface to launch cloud mode.

The heart of this update is "in-place operation." Previously, many cloud-based AI coding tools required developers to create and monitor tasks through a web console. Terminal integration collapses those actions into the same environment where developers already spend their time. For engineers who live in the CLI, reducing context switching means sharper focus and less operational friction.

Four Core Capabilities

The official release outlines four operations available in the terminal, covering the full lifecycle of a cloud session:

Create and Steer

Create lets developers spin up new cloud tasks directly from the terminal without opening a browser. Steer allows users to adjust direction and add instructions while a session is in progress, keeping them in control of the AI's execution path — especially critical during long-running autonomous coding tasks where you need to keep the agent on track.

Resume and Watch

Resume means interrupted sessions can be picked back up in the terminal, which is ideal for complex tasks handled in stages or cross-device collaboration scenarios. Watch provides real-time visibility into a running cloud session, letting developers track the agent's progress from the command line without repeatedly refreshing a webpage.

What This Means for Developer Workflows

This update reflects a broader trend in AI coding tools toward meeting developers in their native environment. The terminal has always been the engineer's primary workspace. Bringing cloud AI session control into the CLI fundamentally transforms Devin from a standalone product interface into a composable building block within an existing toolchain.

The combination of cloud execution and local control is also worth noting. Running sessions in the cloud means compute-heavy, long-running tasks don't tie up your local machine, while the terminal as a control plane preserves the familiar interaction patterns developers already know. This "cloud compute + local control" model is becoming a mainstream architectural pattern for autonomous coding agents.

The "composable building block" framing draws on a long-standing strength of the CLI ecosystem: terminal commands naturally support pipes, scripting, and composition with other tools. Exposing Devin session management as a CLI interface means developers can embed it in Makefiles, CI scripts, or custom shell workflows — not just manual interactions. This follows the same path taken by GitHub CLI (gh) and cloud provider CLI tools like gcloud and aws, where complex cloud operations are distilled into scriptable atomic commands with real automation potential.

Summary

Though brief, this announcement sends a clear signal: Devin is lowering the barrier to cloud AI-assisted coding by consolidating the entire session management loop — create, steer, resume, watch — into the terminal. For developers who rely heavily on the command line, devin --cloud and /cloud make AI-assisted coding a more natural part of their everyday workflow.