Ona Enterprise

Ona Enterprise  is used at Roche as a secure cloud development and AI software engineering environment, where the main use case is to provide a standardized development environment and experiment with AI-assisted software engineering workflows in a controlled enterprise setup. In addition to this, it is also useful for tasks such as setting up development environments quickly, working with repositories in isolated workspaces, running tests and tools in the cloud, using AI agents, as well as coding assistance, refactoring, documentation, and pull request preparations. In an enterprise like Roche, the biggest value is not just the AI coding capability; it is the combination of developer productivity, environment standardization, security controls, and governance.

Ona Enterprise  is usually connected to GitLab  repositories, and then an environment will be automatically set up, after which the cloud coding assistant is used. It also provides better CPU and GPU configurations, so the terminal is usually used to build applications as well as for all other daily development tasks.

The secure cloud setup helps our team because initially, we had a Windows local environment, and sometimes packages break once moved from Windows to Linux. In Ona Enterprise, this is not a concern because there is a Linux terminal that is similar to what we have in AWS , allowing the same thing to be deployed there, making it easy to test it out. Regarding the secure cloud setup, whatever experiments are needed, they will occur in a separate environment in Ona Enterprise. This will not affect anything inside Roche, ensuring that any security breach will not occur in Roche. If any code execution occurs, it will not impact the Roche environment, which helps Roche in terms of security concerns.

The most valuable features Ona Enterprise offers are the secure cloud development environments, as well as background agents and enterprise guardrails. Ona Enterprise also helps reduce local environment issues because the developer and agents can work in a clean, policy-controlled environment. A Linux environment is also provided that is mostly favorable to the production-grade environment, which is similar to what is available in Ona Enterprise as well. The ability to delegate tasks to agents while keeping the work inside an isolated environment is also valuable. This is important in enterprise settings where source code, credentials, and internal systems need stronger controls than what normal local AI coding assistant tools provide. Ona Enterprise also offers SSL support, so it can be connected through local VS Code with SSL.

The main area of improvement for Ona Enterprise is transparency and simplicity around the pricing and usage consumption. Ona Enterprise uses compute units, and while this model makes sense because it combines environment runtime as well as agent usage, it can be difficult for teams to estimate the cost until they have real usage patterns.

Another area of improvement can be onboarding of enterprise users, which was also difficult. Ona Enterprise has many capabilities such as environments, agent automations, guardrails, identity management, and integrations. New users may need clearer guide paths depending on whether they want a basic cloud IDE , AI coding support, background automation, or full enterprise governance.

It would also be beneficial to see more out-of-the-box templates, such as an example of a regulated enterprise environment, especially around safe agent usage, approval workflows, test requirements, and audit practices.

Ona Enterprise has been used at Roche for the last one year.

I have not experienced downtime with Ona Enterprise, but I have experienced issues while connecting with a local SSL server.

Ona Enterprise is designed for scaling assisted software engineering in large organizations. It is especially valuable when many developers or teams need consistent environments, secure access, and controlled use of AI agents. The scalability benefits become more obvious as the number of repositories, developer environments, and AI-assisted workflows increases. For a single small project, it may feel like more infrastructure than necessary, but for an enterprise environment, the standardization and governance are the key advantages.

No other solution was used before Ona Enterprise.

Initially, for every development environment, a better local configuration had to be provided. For example, if a virtual VDI  was being provided for multiple team members, it was necessary to decide what configurations to provide such as RAM, CPUs, and other components, and whether GPU was required based on those needs. After implementing Ona Enterprise, local machines no longer require this level of care because it runs on the cloud. It is no longer necessary to worry about what exactly the local configuration is because every configuration is being utilized by Ona Enterprise.

Before choosing Ona Enterprise, comparisons were made with GitHub  Codespaces , GitHub  Copilot Workspace, coding agents, Cursor , and Vercel. After comparing these options with simple AI coding assistance, Ona Enterprise is more enterprise platform oriented because it is not only about code completion. Ona Enterprise combines cloud development environments with CPUs and GPUs, background agents, policy enforcement, auditability, and enterprise deployment options.

The built-in AI agent in Ona Enterprise already has guardrails. If there is an attempt to access security-sensitive features, it does not allow that. Additionally, when the AI agent needs to execute commands, it is very careful about commands such as delete operations. The platform is reliable for standardized cloud development environments and AI-assisted development workflows. The ability to create clean, isolated environments reduces many local machine problems, so there is no need to worry about local configurations because Ona Enterprise already provides much more configuration at the time of setup. For agentic workflows, human review, automated tests, and clear guardrails are still recommended so that the agent can accelerate work but should not be treated as a replacement for engineering quality controls.

Time management has definitely improved. There is no need to set up the environment every time, which makes development easier. Ona Enterprise automatically clones the repository and has a coding assistant that saves time in daily development purposes.

Ona Enterprise's pricing model is based on compute units, which encompasses whatever compute engines are used for CPU, RAM, and GPU. This covers both environment runtime and agent conversations, which is logical because the platform is not just an IDE  but also runs cloud infrastructure and AI agent workloads. However, the model requires monitoring because usage can vary depending on how long environments run, how many parallel tasks are delegated to agents, and how much automation is used. For enterprise users, it is recommended to start with a controlled pilot, measure compute unit consumption, and then estimate the cost per team or per workflow. For large workflows, the cost can be justified if it reduces developer setup time, improves onboarding, and enables secure AI-assisted development. For a small team or lightweight use case, the value needs to be compared against similar tools.

Ona Enterprise would be recommended for other organizations, particularly large-scale organizations, because it is a strong fit when the organization has multiple repositories, complex development, and strict security requirements as well as needs a standardized developer workflow. It is very important to start with a focused pilot, choose a few repositories, and define the development container setup properly. Overall, I would rate this product a nine out of ten.

On-premises

Amazon Web Services  (AWS)

What do you like best about the product?

Gitpod provides very fast VMs for using our codespaces on very end machines. It has allmost all of our Git platforms to connect and start coding with them. It has many code editors to launch their high end codespaces on them for our development. It also supports browser version of code editors to directly code through any browser and from anywhere we want.

What do you dislike about the product?

Its charges a bit high for limited time of VMs usage.

What problems is the product solving and how is that benefiting you?

Gitpod has provided very high end VMs for our coding, which helps us launch very resource intensive code editors on any browser. It easily connects with all of our git providers, which helps us keep our code updated and managed from anywhere. It has all the code editors we want to code on, which makes our overall development process very simpler and easier.

What do you like best about the product?

Gitpod transforms our development workflow, making onboarding a breeze, enabling secure data access for BYOD setups without the need for a VPN, and enforcing best practices. Most importantly, it turbocharges engineer efficiency. Our engineers can be productive and deploy to production from day one and don't lose time with inconsistent environments. As a result, I (platform engineer) don't have to offer as much support, this having more focus time myself.

The initial dedicated SAAS setup, was up and running in no time. And the support we receive troughout the process has been tremendous.

What do you dislike about the product?

I am an intelij user and setting up the integration wasn't as easy as it was for VS code.

What problems is the product solving and how is that benefiting you?

- Slow onboarding
- No Data access for BYOD
- Inconsistent development environments

What do you like best about the product?

* Onboarding people is a matter of granting them access to gitlab repos, as soon as they have that, they can start coding. No need to deal with Windows laptops where you need admin for everything :D
* One unexpected benefit was that the environments are not only reproducible, they're also isolated: you can work on 2 branches at the same time by spinning up 2 gitpod workspaces.
* Prebuilds ease the pain of waiting for workspace boot.
* We could shut down our previous notebook solution thanks to gitpod using a vscode notebook plugin.

What do you dislike about the product?

The web IDE is close, but not quite the same as regular desktop experience, for example copy pasting shortcuts don't seem to work and of course there's some workspace boot time.

What problems is the product solving and how is that benefiting you?

Onboarding users, reproducible environments (config in the repo itself). It benefits me personally because as a platform engineer, I have to spend less time configuring users' environments. Just click the gitpod button and code away.

What do you like best about the product?

Gitpod allows engineers to focus as much of their time as possible on delivering business value. With Gitpod, engineers don't have to worry about installing tools and can start coding right away. Gitpod workspaces boot up in seconds, run in the cloud (right next to your data) and thus also give you the power of cloud machines. Big enterprises typically don't gave very good engineering laptops, which is also solved by using a Cloud IDE such as Gitpod.

As an internal developer platform team, it allows us to onboard new engineers much faster than before. Where in the past people had to go through 20 documentation pages with all the tools that needed to be installed, this can now be provided by the platform team as a script in the .gitpod.yml configuration file.

What do you dislike about the product?

In general you see that the market is still quite young for cloud IDEs and that you can't use all the features of Jetbrain products yet. Next to that, plugins on the marketplace sometimes don't work.

What problems is the product solving and how is that benefiting you?

- Slow enterprise laptops
- Secure connectivty to data systems (data doens't have to be on personal laptops anymore)
- Onboarding people instantly without losing time
- Helping to bridge the disconnect between the "experimentation" and "industrialization" phases of the data product lifecycle. People can now run Jupyter notebooks right next to their code in an IDE. This promotes better writing of functions and prevents copy pasting code after the machine learning model experiments are done. In the past, when people finished developed a machine learning model in a notebook, they had to go to the next step, which is industrializing the code and making sure the predictions are run everyday or in real-time. People often had to copy paste code from a notebook to an IDE.