New AI assistant in Medusa Cloud

AI coding agents are changing how software gets built. They help teams produce more code, faster, but they also change what happens after that code reaches production.

As more code ships with limited human review, hosting platforms need to take on more responsibility for reliability, debugging, and operational quality. That is the gap we are building Medusa Cloud around.

Medusa Cloud is the operational layer for AI-native commerce development: not just hosting for Medusa applications, but an environment for building and operating them as they scale.

Today, we are taking another step in that direction with a new agent embedded directly in the Medusa Cloud dashboard. It investigates your projects and debugs failing builds and deployments for you.

How it works

In this first iteration, the agent is focused on solving failed builds and deployments.

When a build fails or a deployment breaks, the hard part is rarely seeing that something went wrong. The hard part is collecting the right context, reading the relevant logs, connecting them to the code and configuration, and understanding what actually caused the failure.

That is what the agent is designed to do.

On the deployment details page in the Medusa Cloud dashboard, you'll see a new "Fix with AI" button, if the build or deployment failed:

When you ask about a failing build or broken deployment, the agent retrieves the relevant project data directly from Medusa Cloud and your connected repository. The agent then guides you through the steps needed to resolve the issue with your build or deployment.

This first iteration of our support agent is only the foundation. We plan to expand it in the near term with coding capabilities, enabling you to build features using natural language directly from the Medusa Cloud dashboard.

The agent is available on all plans in Medusa Cloud today. Open any project and start a chat to debug a failing build or deployment.