
What I like most about Daytona is that it lets me run AI-generated code quickly and safely. Sandboxes can be created and launched in just a few clicks, so I can experiment and run code without spending much time or compromising the security of my own machine or servers. They also help me isolate risk because each sandbox is independent, so AI-generated code that negatively impacts one sandbox can’t affect the others.
The Python and TypeScript SDKs make it easy to create, run, and destroy sandboxes programmatically. Snapshots also let me resume a sandbox exactly where I left off, instead of having to recreate everything from scratch. Overall, Daytona reduces the time I spend setting up sandboxes and increases my productivity when working with AI agents. Review collected by and hosted on G2.com.
Daytona has a bit of a learning curve for new users. Concepts like sandboxes, snapshots, and runners could use clearer explanations. I’d also love to see more examples and step-by-step instructions in the documentation, especially for the most common AI agent use cases.
Also, when you’re running many sandboxes, the cost can add up quickly. Usage estimates in the dashboard, along with more controls to manage or limit costs, would be really helpful. Finally, some of the more advanced concepts weren’t available yet (for example, forking a sandbox to run processes in parallel) or were still in development, so I’m looking forward to seeing those features implemented in the future. Review collected by and hosted on G2.com.