kubectl-ai is an open-source, AI-powered Kubernetes assistant from Google Cloud Platform. You describe what you want in plain language — investigate a failing pod, explain a service, scale a deployment — and the agent inspects your cluster, reasons about what it finds, and proposes or runs the kubectl operations to fix it. The project has earned about 7,600 stars on GitHub and is actively developed, with updates in 2026.
The tool works as an agentic loop: it takes your goal, gathers cluster state with kubectl, feeds the results to a large language model, and iterates — checking logs, events and resource definitions — until it can answer or complete the task. It supports multiple model backends, including Google’s Gemini models and other providers, so teams can run it with the AI platform they already use. Interactive sessions let you keep asking follow-up questions about the same cluster, and the agent explains its findings in plain language instead of dumping raw YAML. For platform teams, it doubles as a way to make Kubernetes expertise more accessible: junior engineers can diagnose common failures conversationally, while experienced operators use it to move faster through routine investigations with a full audit trail of what was checked and changed.
kubectl-ai is free and open source under the Apache 2.0 license. You pay only for the LLM API usage of whichever model backend you configure.
kubectl-ai is for Kubernetes operators, SREs and platform teams who want to diagnose and operate clusters conversationally, with an agent doing the kubectl legwork.