Subject and setting
Who or what is visible, plus the environment around it.
Drop a JPG, PNG, or WebP here, up to 10 MB.
The result keeps the parts an image model can act on, without guessing who a person is.
Editorial photograph of a young woman in a structured coral jacket holding a ceramic cup beside a rain-speckled glasshouse window at blue hour. She sits on the right third, looking toward a softly blurred city skyline. Cool blue ambient light balances a warm table lamp, with natural skin texture, subtle window reflections, a 50mm perspective, and shallow depth of field. Quiet, thoughtful mood with a charcoal, blue, and muted coral palette.
The tool turns visible evidence into the details that matter when you generate a new image.
Who or what is visible, plus the environment around it.
Framing, perspective, depth of field, and likely lens cues.
Direction, softness, contrast, and the dominant palette.
The visual medium, texture, atmosphere, and overall feeling.
One focused workflow from reference image to editable prompt.
Use a photo, artwork, product shot, or design reference.
The model checks safety, then describes the visible creative details.
Copy the prompt, change any detail, and use it in your preferred generator.
Use the tool when the image is clearer than the words in your head.
Capture lighting, texture, palette, and camera language from a reference.
Turn framing and spatial relationships into language you can reuse.
Keep the useful art direction, then replace the subject, setting, or mood.
See how a strong prompt connects concrete visual details.
The practical details before you upload.
It describes visible subjects, composition, setting, light, color, camera cues, materials, mood, and visual style. It does not identify real people or infer private traits.
Yes. You can upload a reference from this page, then sign in before Lets Gen analyzes it and creates the prompt.
The tool accepts JPG, PNG, and WebP images up to 10 MB.
Yes. The result is plain editable text, so you can adapt it for Lets Gen or another image model.
The model checks the image as part of the same analysis request. If the image is unsafe or inappropriate, the tool refuses to create a prompt.
The landing page keeps the image only in this browser for up to 15 minutes. It uploads after sign-in as a hidden temporary input and is not added to My Assets.
Move the result into a focused image workflow.