Skip to main content
The prompt box is the entry point to a chat. Use it to explain what you need, add relevant context, choose how you want to interact with Replit, and name the outcome you want.
Workspace permissions, integrations, and security are managed outside the prompt box. The prompt box can use context that is available to you.

​
Ask or describe the task

Write a question, instruction, or task in plain language. You can continue in the same chat as you clarify your request or refine the result. Start a chat to try the prompt box.

​
Add the right context

Give Replit the information it needs for the task. You can attach a file, refer to a connected tool, ask Replit to search the web, or bring in existing work such as a Figma design. Access to connected services depends on your Workspace settings and permissions.

​
Choose how to interact

Choose the experience that fits your task. Agent Modes change how Replit approaches the work. When available, the Model selector lets you choose a model instead of using Auto. Voice Mode lets you speak your request.

​
Name the outcome

Describe the result you want, such as an answer, a recurring Routine, an app, a presentation, a design, or an asset. Replit uses that intended outcome to guide the work that follows.