Nimbus is the cloud engineer agent for AWS and GCP. Connect your cloud and just talk to it. It designs infrastructure on a canvas, diagnoses incidents from real telemetry, tracks spend and finds savings, and ships code fixes as pull requests, all with your approval. Every action runs in an isolated sandbox you review before anything changes.

Hey everyone 👋 I built Nimbus because managing cloud infrastructure still means jumping between the AWS console, Terraform, dashboards, and a dozen CLIs, even for simple things like "why is this API throwing 502s?" or "what am I overpaying for?" Nimbus is a cloud engineer agent you just talk to. Connect AWS or GCP and it reads your real inventory, then you can ask it to design a stack on a canvas, diagnose a live incident from actual telemetry and logs, rightsize resources to cut cost, or fix a bug and open a PR. It reads before it writes and always shows you a plan first, so nothing changes without your approval, and every action runs in a disposable sandbox. The hardest part was making it safe to give an agent real cloud access: per-user credential isolation, temporary keys that expire, and human-in-the-loop approval on every write. Would love your feedback, especially on what you'd want it to handle next. Happy to answer anything! 🙌

How does the sandbox setup actually work behind the scenes, like do I need to provision anything or does Nimbus handle the isolation entirely on its end?

Nimbus

The cloud engineer agent that runs your AWS & GCP

The cloud engineer agent that runs your AWS & GCP