Pick a replacement for a cloud coding environment by the job you need done: a browser editor for yourself, a shared cloud workspace for a team, or a sandbox that runs AI-generated code safely. Those are different products. Shortlist two from the right group, then open one real repository in each before you move a whole team across.
CodeSandbox did several jobs at once, which makes replacing it harder than it looks. Some people used it as a quick browser IDE, some as a cloud workspace for a whole team and some as a safe place to run code they didn't write. Below are nine options, in no particular order, grouped by those jobs.
Which of the four jobs are you replacing?
- A shared cloud workspace. Every developer gets the same configured environment, so new starters can open a pull request on day one.
- A self-hosted or open-source environment. The same idea, but on infrastructure you control, which matters if your code can't leave your own cloud.
- A sandbox for AI-generated code. An isolated box where code written by an agent or a model can run without touching anything important.
- Quick online coding. A browser tab where you try an idea, share a snippet or teach someone, with nothing to install.
Most teams need one of these, and a few need two. Very few need all four, and no single product on this list tries to cover them all.
Workspaces for teams on GitHub, JetBrains or their own cloud
GitHub Codespaces is the obvious starting point if your code already lives on GitHub. It gives you fully configured, secure cloud environments native to the platform, and the vendor says you can start coding in seconds from any device.
JetBrains Remote Development takes a different approach. Rather than a browser editor, it connects your JetBrains IDE to a backend running on a remote machine, which suits developers who don't want to give up the editor they know.
If your environments must run on your own infrastructure, two tools aim squarely at that. Coder provides self-hosted infrastructure for developers and AI coding agents in governed environments. Okteto gives each developer or agent task an isolated, production-like environment, again on your own infrastructure. Both are platform-team purchases rather than something one developer signs up for over lunch.
Open-source, agent sandboxes and browser coding
DevPod is the open-source, client-only choice. It builds dev environments as code from a devcontainer.json file, and because there's no server to run, it's an easy experiment for a team that already uses dev containers.
Daytona is built for a newer job: running AI-generated code inside sandboxes that you create through an SDK. If you're building a product where an AI agent writes and runs code, this is the category to look at, not a classic IDE.
For quick coding in a browser, the two options sit at opposite ends. Replit is now a browser-based builder where you describe what you want and it builds an app or website, which is a long way from a plain editor. OneCompiler stays simple: write, run and share code online, with nothing else attached.
The nine compared
| Tool | Category | Good for |
|---|---|---|
| Coder | Self-hosted cloud workspaces | Governed environments for developers and AI agents |
| Daytona | Sandboxes for AI code | Running AI-generated code in SDK-created sandboxes |
| DevPod | Open-source dev environments | Dev environments as code from devcontainer.json |
| GitHub Codespaces | Cloud dev environments inside GitHub | Configured cloud environments for GitHub repositories |
| JetBrains Remote Development | Remote IDE backend | Using a JetBrains IDE on a remote machine |
| Okteto | Production-like dev environments | Isolated, production-like environments per task |
| Ona | Background agent platform | AI software engineers working from task to pull request |
| OneCompiler | Quick online coding | Writing, running and sharing snippets in a browser |
| Replit | Browser IDE and AI builder | Describing an app and having it built in the browser |
Short notes, A to Z
Coder
Best for: Engineering teams who want to run developers and AI coding agents in governed environments on their own infrastructure, with full control over agent permissions.
Daytona
Best for: Developers building AI products that need to run AI-generated code in secure sandboxes, created on demand through an SDK.
DevPod
Best for: People who want open-source, client-only dev environments defined as code with devcontainer.json; the vendor says there's no server and no vendor lock-in.
GitHub Codespaces
Best for: Teams whose code already sits on GitHub and who want fully configured, secure cloud development environments that start from the repository itself.
JetBrains Remote Development
Best for: Developers who prefer JetBrains IDEs and want to connect to a remote machine running the IDE backend from anywhere.
Okteto
Best for: Platform teams wanting each developer or agent task to get an isolated, full-stack environment that resembles production, running on their own infrastructure.
Ona
Best for: Teams wanting AI software engineers to run in the cloud in the background, taking a task and returning a pull request, as the vendor puts it.
OneCompiler
Best for: Students and developers who just need a free online compiler to write, run and share code in the browser.
Want AI doing the heavy lifting in your marketing?
I build the systems that handle the boring 80 percent, so you get your week back. Done properly, with the human kept in.
Replit
Best for: Non-developers and makers who'd rather describe what they want and have an app or website built for them in a browser.
Names that have moved or closed
This corner of developer tooling is losing names, so check a tool's own site before you trust any older comparison. At the time of writing, Codeanywhere's own site says it's sunsetting, and Glitch's blog carries a farewell post. The old Gitpod address now redirects to Ona, a product that pitches background AI agents taking a task and returning a pull request, rather than the cloud workspaces Gitpod was known for.
None of that is a reason to avoid the category. It's a reason to pick tools whose direction matches yours. If you want human developers in cloud workspaces, choose a vendor still building for that. If you want agents doing background work, the newer agent platforms are where the effort's going.
Open-source options soften the risk too. If a project's maintainers move on, the code you've got keeps working while you decide what's next.
My advice for engineering leads
Start from where your code and your security rules already live. If your repositories are on GitHub and your code can sit in a vendor's cloud, the native option will be the least work. If your code has to stay inside your own infrastructure, look at the self-hosted platforms and plan for the platform team's time to run them.
Keep AI sandboxes and developer workspaces as separate decisions. A sandbox for agent-written code is a product feature you're building for customers, while a workspace is a tool for your own team. Mixing the two in one evaluation tends to produce a choice that does neither job well.
If what you really miss is a quick browser editor, don't buy a team platform. A simple online compiler or a browser builder will do. You'll find more neighbours in my round-up of browser IDE alternatives and my list of front-end hosting alternatives.
A one-repository trial
- Pick two tools from the same job group on this page.
- Choose one real repository your team works on every week, not a starter template.
- Open it in both trials and time how long it takes until tests run green.
- Ask a newer team member to make a small change and open a pull request without help.
- Check where the environment runs, who can see the code and how access is removed when someone leaves.
- Close both environments and reopen them the next day to see what was kept.
If both tools pass, choose the one your team grumbles about least. That's the one they'll keep using.
Do you make a cloud dev environment or code sandbox? Tool makers can ask to be considered for a featured place on this list.
Want your tool on this list? Featured spots and reviews here are paid partnerships. Here's how to partner with me.
Frequently asked questions
What's a cloud development environment?
It's a ready-configured coding setup that runs on a remote machine instead of your laptop. You open it in a browser or connect your editor, and everyone on the team gets the same tools and settings.
Do I need a sandbox for AI-generated code?
If your product lets an AI agent write and run code, yes. A sandbox keeps that code isolated, so a mistake or a bad instruction can't reach your other systems or your customers' data.
Is a self-hosted environment worth the effort?
It's worth it when your code or data can't leave your own cloud, or when you need tight control over who and what can run there. For a small team without those rules, a hosted option is usually less work.
What's devcontainer.json?
It's a configuration file that describes a development environment as code: the base image, tools and settings. Several tools can read it, so it makes moving between environment providers easier.
How do I know if a coding tool is being wound down?
Check the vendor's own site and blog, and see whether the address still loads or redirects somewhere else. Old reviews won't tell you, because they were written before any change.
