Five years ago we wrote about using Docker for local development the right way. A lot has changed since then — except our obsession with developer experience. Today, we're entering a new chapter.
Back in 2021, Kool was born out of frustration. We were a software house juggling dozens of client projects across different stacks, and Docker — while powerful — kept getting in the way. Permission issues on mapped volumes, MTU mismatches, the -t vs -T flag dance, teammates siloing Docker knowledge... we had seen it all.
So we built Kool: a CLI tool that wrapped Docker and Docker Compose with sane defaults, presets for popular frameworks, and custom images optimized for local development. The goal was simple — remove the friction so developers could focus on code, not infrastructure.
The developer community responded. Our articles on Dev.to reached over 33,000 views. We went from 60 to 600+ GitHub stars. People joined our Slack, opened issues, contributed code. Kool had found its audience.
Five years later, the problems Kool was built to solve haven't gone away. If anything, they've gotten worse. Docker's ecosystem has grown more complex. Multi-service architectures are the norm. Teams are more distributed. Onboarding new developers onto a project still involves too much "works on my machine" debugging.
The original arguments for Kool remain exactly as strong today:
kool.yml so every team member runs the same commands.docker-compose.yml is yours. Kool wraps it, never replaces it. You keep full control.kool create still spins up a fully configured local environment in minutes. kool run still keeps your workflow scripts shared and versioned. The original value proposition holds stronger than ever — because the fundamental problems haven't changed, and neither has Kool's answer to them.
In 2023, we wrote about our journey and the natural next step: taking the same Docker Compose setup developers use locally and deploying it to the cloud. In early 2024, we launched Kool Cloud publicly.
We spent nearly two years running and refining the platform with a curated group of early adopters. That deliberate pace taught us what works in production — and what developers actually need when moving containers to the cloud. We iterated on the deploy experience, built out managed MySQL databases, environment management, team collaboration, and a billing model that makes sense for small teams and growing companies alike.
The cloud platform matured. Quietly, intentionally, and with real users validating every decision.
Here's what's different now: AI agents are writing code faster than teams can ship it.
Coding agents — Claude, Copilot, Cursor — are generating features, fixing bugs, and refactoring entire codebases in minutes. The bottleneck is no longer writing code. It's everything that happens after: spinning up a preview environment for PR validation, handing off to QA with a live URL, configuring deploys, provisioning infrastructure, debugging Kubernetes manifests, understanding cloud provider quirks. The same learning curve problem we solved for Docker locally now exists for the entire path from code to production — and it's worse, because the speed gap between code generation and code shipping has never been wider.
Developers still need to feel in control. But the cost of that control should be managed. You shouldn't need to deep-dive into specific DevOps toolchains just to get your app live. Not when AI just wrote it in 30 seconds.
This is where Kool's new chapter begins.
We're building Kool to be the deployment layer that AI agents interact with natively. Your coding agent already writes your code — now let it deploy too. Deploy your app, provision a database, check logs, roll back — agents can do all of this on their own, autonomously, as part of their workflow.
But here's the thing: the same interface works just as well for you. The Kool CLI was always designed for low learning curve, and that hasn't changed. kool cloud deploy is one command. You don't need to understand Kubernetes, Helm charts, or cloud provider consoles. Whether it's your agent running the command or you typing it yourself, the experience is the same.
Optimized for agents. Accessible to humans. No DevOps degree required for either.
Two paths, one experience:
Managed Cloud — Deploy your Docker apps to Kool Cloud in minutes. No Kubernetes knowledge required. No servers to manage. Just kool cloud deploy and your app is live on infrastructure managed by us. Think of it as borrowing a sane, simple cloud management layer so you can focus on your product. You pay a premium for the convenience — and that's a fair trade when you're getting started or want zero ops overhead.
BYOC (Bring Your Own Cloud) — When you want ownership and cost efficiency, bring your own servers. Cherry-pick your VPS provider, choose your architecture (amd64 or arm — yes, Graviton works), right-size your hardware for your workload. Kool installs and manages the entire stack on your servers — same deploy experience, same dashboard, same agent integration. You own the hardware, you control the costs, we provide the intelligence to manage it.
With BYOC, the savings are real. You're not paying managed cloud margins on top of compute you could provision yourself for a fraction of the price. Pick Hetzner, DigitalOcean, AWS, or any provider you like. Run ARM instances where it makes sense. The infrastructure choices are yours — Kool just makes sure deploying to it stays simple.
The transition is seamless. Same CLI. Same dashboard. Same deploy workflow. Same AI agent commands. The only thing that changes is where your containers run — and how much you pay for them.
Start by borrowing. Graduate to owning. No re-platforming required.
The learning curve argument that made Kool relevant in 2021 is stronger now than ever. Back then, Docker complexity was the bottleneck between an idea and a working local environment. Today, cloud deployment complexity is the bottleneck between AI-generated code and a live product.
We're building toward a future where:
Kool started as a tool to make Docker suck less locally. It's becoming the platform that makes deploying suck less too — with AI agents as the interface and your choice of infrastructure underneath.
We're just getting started on this new chapter. If you've been following Kool's journey since the early days, thank you — the best is ahead. If you're discovering us for the first time, there's never been a better moment to give Kool a try.
Follow our journey:
Running web applications on containers in the cloud is a breeze with Kool.dev Cloud. Easily move any Docker Compose based environments to our managed cloud powered by Kubernetes (AWS EKS).
Checkout the highlights of what was shipped in the latest Kool release 3.0.0. Interactive cloud deploy configuration helper, preset updates, and much more.
We are thrilled to announce that Kool.dev Cloud is now generally available! Join today and be a part of the excitement!
