The open testing platform that took your tests out of the pipeline - and put them where your software actually lives.
Somewhere right now, a deployment is shipping. A team merged code an hour ago, an AI assistant wrote half of it, and the cluster is humming along. The tests? They are not stuck in a queue behind a build agent. They are running inside the same Kubernetes environment the app runs in - in parallel, isolated, triggered by an event nobody had to babysit. That quiet competence is the thing Testkube sells.
Testkube is a Kubernetes-native test orchestration and execution platform. Translated out of conference-speak: it runs any test, with any tool, at scale, inside your own clusters - and then collects every result, log and artifact into one place. It is roughly 27 people spread across the Americas, Europe and Asia, an MIT-licensed open-source engine on GitHub, and a commercial control plane on top. Modest headcount. Immodest test count.
The category it occupies barely existed a few years ago. "Test orchestration for Kubernetes" was the kind of phrase that earned blank stares at conferences. Today it describes a real and growing need: the more of your stack runs in containers, the stranger it feels to test that stack from somewhere outside the cluster. Testkube's wager is that this awkwardness is permanent, and that someone should build the tool that resolves it properly rather than asking every team to improvise.
Here is the uncomfortable part. Most testing tools were designed back when an application was a thing you installed on a server, not a swarm of microservices breathing in and out of a cluster. So teams did what teams do: they bolted tests onto the CI/CD pipeline, where they slowed builds, fought over runners, and conveniently fell apart the moment anyone tried to test a service that only behaves correctly inside Kubernetes.
The result was a pile of homegrown scripts, brittle glue code, and a dashboard for every tool. Each team reinvented the same wheel, badly, and called it a testing strategy. It worked - in the same way a parachute made of bedsheets technically works.
Then AI showed up and made it worse. When machines write code faster than humans can review it, the bottleneck moves. It is no longer how fast you can write software. It is how fast you can trust it.
Testkube was founded in 2023, spun out of the cloud-native studio Kubeshop. Its co-founder and CTO is Ole Lensmar, the creator of SoapUI - one of the most widely used API testing tools ever shipped. He had already watched an entire generation of engineers learn to test APIs his way. The bet this time was narrower and bolder: stop treating testing as an afterthought stapled to the pipeline, and rebuild it as something native to the place software actually runs.
Alongside him, Dmitry Fonarev runs the company as CEO. The wager underneath both of them is almost philosophical. If testing lives inside Kubernetes - not beside it, not after it - then it can scale the same way everything else in the cloud-native world scales. Horizontally, on demand, and without a human holding the door open.
The core move is deceptively simple: decouple testing from CI/CD. Instead of asking your build server to also be your test runner, Testkube lets you define tests as code and trigger them however you like - on a schedule, from a pipeline, from a Kubernetes event, over the REST API, or via an MCP server so an AI agent can launch your suite directly. The tests run in the cluster, in parallel, isolated from one another.
API, end-to-end, performance, security or infrastructure tests - using the tools you already know, at scale.
Schedules, CI/CD, Kubernetes events, REST API or MCP. Even an AI agent can kick off a run.
Results, artifacts, logs and resource metrics, aggregated into a single dashboard across teams and environments.
AI-powered troubleshooting that helps teams diagnose failures instead of spelunking through logs by hand.
For the enterprise crowd there is the unglamorous-but-essential layer too: SSO and SCIM, role-based access control, audit logging, and multi-cluster management. The parts nobody tweets about and every security review demands.
What ties it together is reuse. Testkube does not ask you to abandon the testing tools your team already trusts - k6 for load, Postman for APIs, Cypress and Playwright for the browser, JMeter for the old-school crowd. It wraps them, schedules them, isolates them, and hands you a consolidated view of what passed and what did not. The pitch to an engineering leader is refreshingly free of drama: keep your tools, keep your infrastructure, and stop paying the tax of gluing them together by hand.
Skepticism is the correct posture for any developer-tools pitch, so here is the evidence. Testkube has powered more than 100 million automated tests on customer infrastructure. Its customer list is not a roster of logos hunting for a free trial - it includes Cloudflare, Siemens, Adobe and Volvo, organizations with little appetite for flaky testing.
Roughly 51 million tests by June 2025, doubling past 100 million within the same year. The bars are approximate, the trend is not subtle.
The money arrived in September 2025: an $8 million Series A led by Insight Partners and Ratmir Timashev, the co-founder of Veeam. Backers who built infrastructure businesses tend to recognize one when they see it. The round was earmarked for the obvious things - expanding the platform, growing the team - and one less obvious thing: keeping testing reliable as development velocity climbs faster than any human reviewer can follow.
There is a tidy symmetry in the cap table. Veeam turned the dull, essential work of backup and recovery into a multi-billion-dollar company by being relentlessly good at the part everyone else found boring. Testkube is making a similar argument about testing. Nobody dreams of a career in test orchestration. Everyone, eventually, needs it to work.
Stripped of jargon, Testkube's mission is to help software teams ship faster with better quality and lower cost - by unifying the fragmented mess of test automation into a single Kubernetes-native platform. It is a remote-first, open-source-rooted company that lives inside the cloud-native community it serves, which is a polite way of saying its users and its engineers tend to argue in the same Slack channels.
Return to that quiet deployment from the opening. The reason nobody had to babysit it is the whole point. As AI generates more of the world's software, the volume of code outruns any team's ability to manually verify it. Continuous, automated, scalable testing stops being a nice-to-have and becomes the safety net that lets you trust output you did not personally write.
Testkube's bet is that the net has to live where the software lives - in the cluster, native, always on. The pipeline was never the right home for it. That is the change Testkube is trying to make ordinary: a deployment that ships, tests itself thoroughly, and never once asks a human to hold the parachute.