Image to Prompt is here

Attach up to three images to an idea and turn what is visible into a detailed, reusable prompt.

Image to Prompt now lets you attach JPEG, PNG, or WebP references alongside a typed or spoken idea. PromptDock uses a vision-capable model, asks focused follow-ups only when intent is unclear, and creates a self-contained prompt you can edit, copy, save, or share. Groq switches to its vision model automatically; saved providers can use a separate image model from Settings. Images are sent only during generation and are not saved in your library.