General prompt
A natural, detailed description for broad image generation.
Visual intelligence, under your direction
See the image. Keep the intention.
ImagePrompt reads subject, light, composition, materials, palette, and mood — then turns them into a production-ready prompt you can actually direct.
PNG, JPG, WebPUp to 4 MBEditable output
01Bring a referenceUpload, paste, or link
02Read the visual systemSubject, light, palette, lens
03Choose the prompt languageGeneral, structured, model-ready
04Refine your directionEdit, copy, reuse
The visual lab
Drop in a picture. Get more than a caption.
Choose how deeply the image should be interpreted and which creative model or workflow the result should serve.
Reference intelligenceNot keyword guessing.
Why it feels different
It reads relationships, not just objects.
A useful prompt needs more than “flower on a table.” It needs hierarchy, distance, surface, direction of light, camera behavior, atmosphere, and the visual tensions that make the reference memorable.
- 01CompositionFraming, balance, depth, point of view
- 02Material & lightTexture, reflection, softness, direction
- 03Creative intentMood, medium, restraint, model-specific language
From visual reference to structured direction.
A short workflow built around control. You decide the source, the output format, and the amount of detail.
01 / SOURCE
Bring the reference
Upload, paste from your clipboard, or provide a public image URL.
02 / DIRECTION
Choose the lens
Select a general, structured, model-specific, or machine-readable prompt format.
03 / RESULT
Refine and reuse
Edit the generated prompt, copy it, and keep recent versions in your local history.
Visual vocabulary
One tool, radically different creative worlds.
Photography, product design, illustration, architecture, fashion, and surreal concept art all need different descriptive language.
One image. Several ways to describe it.
Different tools expect different prompt structures. The studio adapts the result before you leave the page.
Structured prompt
Subject, composition, light, palette, camera, and atmosphere.
Flux direction
Dense visual language tuned for modern generative models.
Midjourney format
A concise creative prompt with optional parameter guidance.
Stable Diffusion
Positive prompt structure with negative prompt suggestions.
Graphic design
Layout, hierarchy, typography, color system, and medium.
JSON output
Machine-readable fields for automations and creative pipelines.
Multilingual
Generate the result in your preferred output language.
Private by design
Your reference is input, not inventory.
The processing layer is designed to avoid permanent image storage. Request limits and signed server-to-server calls protect the generation endpoint.
No AI credentials exposed in the browserValidated file types and size limitsRate-limited generation requests
Questions, answered clearly.
What image formats can I upload?
The first release accepts PNG, JPEG, and WebP images up to 4 MB.
Can I edit the generated prompt?
Yes. The result is placed in an editable field before you copy or reuse it.
Does the browser receive the AI API key?
No. The browser talks to WordPress, and WordPress signs a protected request to the processing worker.
Can I use a public image URL?
Yes. The server validates and downloads supported public images before analysis.