Build, run, and refine in one place
Open the workspace and choose Build. Describe what you want to make, who it is for, and the behavior it should have. InriGPT can create project files, explain the implementation, and revise the result as the conversation develops.
The editor supports common source files, including HTML, CSS, JavaScript, TypeScript, and Python. Project notes keep requirements and decisions alongside the files so follow-up requests have useful context.
A real cloud terminal for your project
Each project can use a separate Daytona Linux environment. Run commands, install npm or Python packages, execute tests, and inspect the actual output. The terminal is a command runner; long-running websites use the separate Start server action.
For Python projects, create a virtual environment before installing dependencies. For a website, ask for a server that listens on the configured port and exposes an HTTP preview. Review the logs when something fails and ask the AI to fix the reported problem.
Try a concrete first project
“Build a responsive reading tracker with HTML, CSS, and JavaScript. Let me add books, mark them as finished, and save the list locally. Use accessible controls, test the main interactions, and explain how to run it.”
Start with a small feature you can verify. Ask for meaningful tests, try the controls yourself, then request the next improvement. Generated code can contain mistakes, so successful output and your own review matter.
Preview and keep what you build
Use the website preview to check the result. Cloud servers can create signed links that expire when the server or environment stops. Static share links expire after seven days. Anyone with a share link can view the included project.
Download your source as a ZIP when you want to keep it. Remote environments are temporary; installed dependencies and files that exist only in the cloud can disappear after the environment ends. Editor files remain in the browser conversation.