Free ImageFX prompt worksheet · no sign-up
Upload a reference image and separate its subject, scene, medium, light, color and composition before copying one fluent prompt. The result is built for the next step in a Google ImageFX workflow, while every detail stays editable.
Reference image analysis
Editable worksheet
ImageToPrompt.me generates text from the image you submit. It is not affiliated with or endorsed by Google, and it does not generate images through Google ImageFX on this page.
Each card pairs one visible reference with a matching natural-language prompt. These are prompt examples, not claims that the reference image was generated by ImageFX.
A ceramic artist in a charcoal apron shapes a small clay vessel with both hands on a turquoise pottery wheel inside a working studio. Frame the scene from waist level with the hands and wet clay as the focal point, shelves softly blurred behind them, cool natural window light, earthy neutral colors and an honest documentary-photography mood.
Change one thing: older craftsperson · warmer side light · tighter hand close-up
A matte white pump bottle rests diagonally on a pale stone tray beside glossy green leaves and a small piece of driftwood. Use a clean overhead product-photography composition with generous negative space, crisp geometric shadows from strong daylight, restrained white and green colors, tactile natural materials and a fresh editorial skincare mood.
Change one thing: amber glass bottle · soft cloud light · warm limestone surface
Weathered stone ruins stand alone on a rugged grass-covered sea cliff, viewed from below so the broken walls cut into a wide steel-blue sky. Keep the landscape sparse and windswept, with fine mist in the distance, subdued cool colors, soft overcast light, rough rock textures and a quiet cinematic sense of isolation.
Change one thing: dawn fog · lone traveler · winter coastline
A refined Japandi living room with a pale sofa, sculptural wood chair, low dark coffee table and a circular wall mirror faces a large city window. Use a balanced wide interior composition, soft morning daylight, warm beige and charcoal accents, clean architectural lines, natural wood textures and an uncluttered, quietly luxurious editorial atmosphere.
Change one thing: rainy skyline · evening lamp light · more lived-in styling
A crowded city street at night glows with layered red, cyan and white signs while pedestrians move beneath umbrellas on rain-darkened pavement. Use a street-level wide view with converging storefronts, reflections across the foreground, high contrast neon light, deep shadows, crisp urban detail and an energetic documentary mood without inventing readable brand text.
Change one thing: fewer pedestrians · blue-hour rain · low-angle reflection
A small rustic tomato galette with deeply golden folded pastry and glossy roasted cherry tomatoes sits on a simple white plate beside a fork. Frame it close from slightly above, keep the wooden table and background softly blurred, use gentle natural side light, warm appetizing colors, crisp flaky texture and an unfussy farmhouse food-photography mood.
Change one thing: dark ceramic plate · overhead view · summer picnic setting
A generic caption often mixes fixed visual facts with details you intend to change. This worksheet keeps those decisions visible before it assembles the final description.
The generator creates editable text for a separate ImageFX workflow. It does not recover hidden metadata or send your content to Google automatically.
It analyzes visible details in a reference image and creates an editable worksheet plus one natural-language prompt. You can copy that prompt into Google ImageFX and revise the suggested subject, scene or style.
No. ImageToPrompt.me creates the prompt. Google ImageFX is a separate Google Labs product where you can paste the prompt and generate images. No image or prompt is transferred automatically.
No. A finished image does not reveal the full original prompt, model version, seed or editing history. The output is a practical visual description to review and refine, not forensic recovery.
You can upload JPG, PNG and WebP files. Large images are resized in the browser before analysis. The free allowance is shared with the main ImageToPrompt generator.
Break the same reference into short subject, setting, style, light, composition and mood tags, then keep only the details that matter.