Design from a PRD or brief

Upload requirements and have the AI read them before it designs anything.

Last updated

If the thinking already exists in a document, don't retype it into a prompt.

Steps

  1. Open your project and upload the document to the project's documents area (or attach it in the prompt box).
  2. Ask for what you want built from it:

Read the PRD and design the screens it describes. Start with the plan.

  1. The AI lists the project's documents, reads the relevant one, and uses it as the source of truth for screens, content and terminology.

Supported documents

PRDs, design briefs, specs and similar text documents. What matters is that the text is extractable — a scanned image of a document is a picture, not a document, and should be uploaded as an image instead.

Getting the most out of it

  • Name your screens in the document. A PRD with an explicit screen list produces a far better plan than one that only describes behaviour.
  • Point at the section you mean. "Design the screens in section 4 only" beats asking it to design the whole document at once.
  • Keep the document in the project. It stays available for later turns, so months-later follow-ups still have the original requirements.

Cost

Listing and reading documents is free. Only the screens designed from them cost credits.

Was this article helpful?

Related articles

Read this in your language