live demo & free trial

Break something. Watch Krenox explain it. Then trial it on your own estate.

First, trigger a real incident in our sandbox and watch Krenox detect it, assemble the evidence, and walk you to the root cause, live. Then take it home: a free trial period in your own AWS account, full platform, no card, no commitment, ample time to validate it. Stay only if it earns its keep.

Start your free trial See published pricing →

The button opens your mail app with a short questionnaire prefilled. No mail app? Write to hello@krenox.ai from anywhere.

the sandbox · opening soon

Feel an incident before you buy anything

you pick the failure
✓CPU pinned at 99% on a service in the request path.
✓A pod OOM-killed mid-traffic, again and again.
✓Disk filling to the brim on a node that matters.
✓A crash-looping deploy or a latency spike with no obvious cause.
you watch krenox work
◆Detection fires from math, not a model's mood: you see the exact rule that paged.
◆An evidence bundle assembles in front of you: metrics, eBPF socket data, the deploy diff.
◆The AI explains the root cause with receipts, states its confidence, and names its blind spots.
◆Experimental: a proposed fix. Where Krenox is confident, it opens a pull request with the full evidence attached. Nothing ships without a human.

The demo environment is a real EKS estate we run, and it's being wired up now. Leave your email and you'll be first in when it opens.

Get notified when it opens →
the trial period

Then run it where it counts: your estate

what you get
✓A free trial period in your own AWS account: the full platform, not a cut-down tier. No card, no commitment, ample time to validate.
✓White-glove install and tuning for your estate, on us.
✓A direct line to the founders: the people who wrote the code answer your questions and take your pages seriously.
✓Walk away clean, any time. One-command uninstall, your data downloaded from your dashboards, nothing retained.
what we ask
◆A real estate, not a toy cluster. Krenox proves itself on production traffic or not at all.
◆A monthly working session with your SRE/platform team while the trial runs.
◆Honest feedback: what worked, what paged you at 3am, what you'd rip out.
◆Patience with rough edges. You're trying an early product. We'll always tell you which parts are which.
after the trial
The first five companies to convert get ~50% off their first year.
Start your free trial
who it's for

A good fit looks like this

01
50–500+ engineers, on AWS

Enough production surface that reliability is a real cost center, running Kubernetes and/or VMs in your own AWS accounts.

02
Paying metered observability bills

You already know what per-GB pricing feels like in an incident month, and you'd like your bill to stop being a variable.

03
An opinion about AI in production

You want AI in your incident response, but only on terms you can audit. Our human-approval and sovereignty model should feel right, not restrictive.

honest expectations

What early access actually means

The core loop (observe, detect, explain, propose a fix, remember) works today and runs on our own infrastructure. Coverage is deepest on EKS and EC2; some collectors are younger than others, and we'll tell you exactly which before you sign. You'll hit bugs. You'll also get fixes at a speed no large vendor will ever match, from engineers who are embarrassed by every one of them.

The trial runs long enough to hit a real incident month and see what Krenox remembers by the end of it. If it hasn't earned a line in your budget by then, walk away clean: nothing retained, nothing owed.

start your trial
One email. No form, no qualification call script.
Tell us three things:
1rough org size and AWS footprint
2your stack: K8s, VMs, CI/CD
3what your current observability costs you
hello@krenox.ai
A founder replies within two business days, with a straight answer either way.