OpsKit vs k8sgpt
k8sgpt is an Apache-2.0 CLI that scans a cluster and explains what it finds. OpsKit puts the cluster inside your AI coding agent, with evidence contracts and GitOps PRs. Where each one fits.
Facts about k8sgpt taken from github.com/k8sgpt-ai/k8sgpt, checked 29 July 2026. Anything their site doesn't state is marked "not stated" rather than guessed at.
What k8sgpt is
An open-source (Apache 2.0) CLI and in-cluster operator that scans Kubernetes, triages issues, and explains them in plain English using an AI backend of your choice.
What OpsKit is
A Kubernetes and cloud investigation kit for AI coding agents. It adds a read-only MCP connection to your live cluster, investigation skills with explicit evidence requirements, and a safety layer that blocks every mutation verb before it executes.
The short version
k8sgpt is free, mature, and genuinely good at what it does: sweep a cluster and tell you what looks broken. It is a scanner. OpsKit is an investigation workflow inside the agent you already code in — it follows a specific failure, gathers the evidence a senior SRE would gather, and turns the result into a pull request. If you want continuous cluster scanning, use k8sgpt. If you want to work one incident to a cited root cause, that's OpsKit.
Side by side
| Dimension | OpsKit | k8sgpt |
|---|---|---|
| Shape of the tool | A kit installed into your AI coding agent | A standalone CLI, plus an in-cluster operator |
| Cluster access | 15 read-only MCP tools | Read-only — analyses without modifying resources |
| How you use it | Ask about a specific failure; the agent investigates it | Scan the cluster; get a triaged list of issues |
| Coverage | Kubernetes plus AWS, GCP, Azure, Terraform and incident response | Kubernetes analyzers — pods, PVCs, services, ingresses, nodes and more |
| Produces a fix | Generates a GitOps PR with rollback plan and evidence attached | Explains and recommends; does not apply fixes or open PRs |
| Licence and price | Commercial — $19 one-time | Apache 2.0 — free |
Choose k8sgpt if…
- You want continuous, scheduled scanning of a whole cluster rather than a per-incident investigation.
- You need it to run in-cluster as an operator, or inside CI, with no coding agent involved.
- Budget is zero and Apache-2.0 licensing is a requirement.
- You want to pick your own AI backend, including a local model via Ollama or LocalAI.
Choose OpsKit if…
- You already work in Claude Code, Cursor, or Kiro and want the cluster available where you're already typing.
- You want the investigation to state what evidence it used — and abstain when that evidence is missing.
- You want the outcome to be a reviewable pull request against the GitOps repo that owns the resource.
- You need the same workflow to cover AWS, Terraform and incident command, not only Kubernetes objects.
Questions
Do they conflict if both are installed?
+
No. k8sgpt runs as its own binary or operator and OpsKit runs inside your coding agent. Both read the cluster read-only, so running them side by side is safe. Several teams scan with k8sgpt on a schedule and investigate specific findings with OpsKit.
Why pay when k8sgpt is free?
+
You shouldn't, if scanning is what you need — k8sgpt is excellent and free. OpsKit is worth paying for when you want a per-incident investigation protocol with evidence contracts, an enforced mutation block, and a GitOps PR at the end.
Try the free tool first
The Kubernetes YAML linter runs the same class of checks OpsKit's deployment guard applies — free, in your browser, no signup. It's the cheapest way to see how the kit thinks.