Natural Language

Describe the screenshot you want

Skip the parameter docs. Pass a prompt string in plain English and the API resolves it to the right device, background, perspective, shadow, and text. Costs +1 credit on top of the resolved render.

How it works

  • The prompt is sent to Claude, which maps your description to ScreenFramed's parameter schema. Ambiguity is resolved with sensible defaults: "dark" becomes background_preset: "midnight", "iPhone" becomes device: "iphone-16-pro", and so on.
  • Pass show_resolved_params: true to see exactly what the prompt resolved to. Useful for building the prompt once, then switching to explicit parameters for production.

Example prompts

  • "MacBook Pro on a sunset gradient, hero angle, with a float shadow"
  • "Clean documentation screenshot, no device frame, minimal padding"
  • "App Store screenshot for iPhone, portrait, dark background with title 'Track Your Habits'"
  • "Social media square, browser frame, ocean preset, soft shadow"
  • "OG image for a blog post with the title 'Shipping v2'"

API example

curl -X POST https://screenframed.com/v1/capture \
  -H "Authorization: Bearer sf_live_your_key" \
  -H "Content-Type: application/json" \
  -d '{
    "url": "https://your-app.com",
    "prompt": "MacBook Pro on a sunset gradient, hero angle, float shadow",
    "show_resolved_params": true
  }'