How accurate are the prompts?

Subject, framing, light, and color are restored item by item, up to 90 percent on mainstream styles.

Which models are supported?

MJ V7, Flux Pro, Wan video, and universal tags, all paste-ready.

Yes, but only style and lighting descriptions, never real identity features.

Upload a screenshot for Wan-dialect prompts including camera moves.

Bilingual, pure English, and pure Chinese; prompt bodies stay English keywords.

Files stay in your encrypted workspace session, never shared, deletable anytime.

Upload any image for copy-ready prompts with style and lighting decoded.

Upload an image for MJ and Flux-ready prompts with composition and lighting breakdowns. Free online tool. Crea gratis online con Loomi AI.

Dialect-Accurate Output——MJ, Flux, and Wan dialects ready to paste straight into generators.
Lens-Level Teardown——Subject, framing, light, and color itemized into reusable recipes.
One-Click Lookalike——Re-run the recipe for up to 90 percent visual fidelity.

Upload the Image Drop in a JPG, PNG, or WebP file up to 10MB with an optional focus note.
Pick Dialect and Depth Choose model dialect, teardown depth, and output language.
Copy the Recipe and Run Receive copy-ready prompts and tag cards, paste to generate.

How accurate are the prompts?: Subject, framing, light, and color are restored item by item, up to 90 percent on mainstream styles.
Which models are supported?: MJ V7, Flux Pro, Wan video, and universal tags, all paste-ready.
Can it decode portraits?: Yes, but only style and lighting descriptions, never real identity features.
Can video frames work?: Upload a screenshot for Wan-dialect prompts including camera moves.
Which output languages?: Bilingual, pure English, and pure Chinese; prompt bodies stay English keywords.
Are my uploads private?: Files stay in your encrypted workspace session, never shared, deletable anytime.

2026-10-22