Use this existing Omellody utility to turn a rough image idea into a complete prompt package with subject, style, lighting, composition, format, negative prompts, and reusable variations. The builder runs locally in your browser and does not send your inputs to Omellody.
Replace the examples with sanitized project details. The generated prompt updates locally in your browser.
Formula: Subject + use case + style + lighting + composition + material details + output format + negative prompt + variation request.
| Variable | What to enter | Example |
|---|---|---|
{subject} | The main product, scene, person, object, or thumbnail concept. | a reusable insulated water bottle for remote workers on a clean desk |
{style} | The visual direction: photorealistic, editorial, 3D render, minimalist, cinematic, illustrated, and so on. | photorealistic commercial product image |
{lighting} | Light source, contrast, mood, shadow quality, and any reflection constraints. | soft side light, gentle shadow, warm neutral highlights |
{composition} | Framing, camera angle, focal hierarchy, crop, copy space, and background relationship. | centered hero shot, shallow depth of field, uncluttered negative space |
{details} | Textures, props, colors, audience clues, brand constraints, and factual limitations. | matte finish, condensation beads, laptop edge in background, no visible logo, premium everyday feel |
{format} | Aspect ratio, platform, crop, resolution target, and any safe-zone requirements. | 4:5 portrait for social and landing-page hero crops |
{negative} | Defects, misleading elements, rights risks, and visual artifacts to avoid. | blurry, distorted hands, unreadable text, extra bottles, watermark, trademark logo, low resolution |
| Goal | Best starting point | Extra variable to add |
|---|---|---|
| Landing-page hero | Product hero or scene-setting prompt with strong copy space. | Above-the-fold crop, CTA-safe blank area, and brand palette. |
| Ad creative | Conversion-focused variation with one visual promise and one audience cue. | Hook placement, mobile crop, compliance-safe claim wording. |
| Editorial image | Documentary or magazine-style prompt with factual context. | Source context, realism level, and no misleading reconstruction. |
| Thumbnail | High-contrast composition with one focal subject and readable empty space. | Small-screen legibility, face/object scale, and background simplicity. |
| Model family | Keep from this generator | Adapt after generation |
|---|---|---|
| Midjourney-style tools | Subject, style, composition, lighting, aspect ratio intent. | Move ratio and stylization into model flags if supported. |
| Stable Diffusion-style tools | Main prompt, detailed prompt, negative prompt, texture constraints. | Separate positive and negative fields, then tune sampler/steps outside the prompt. |
| DALL-E / Gemini-style tools | Plain-language brief, rights-safe descriptors, output use case. | Use shorter natural language and avoid overloading with parameter syntax. |
| Design workflow handoff | Detailed prompt, review checklist, channel constraints. | Convert model output into a human art-direction note before publishing. |
Before using an output, score it with this lightweight formula: Prompt score = subject accuracy + channel fit + composition clarity + rights safety + defect control. Use a 0-2 score for each dimension; publish only results that reach 8/10 and have no rights or privacy failure.
| Score dimension | 0 | 1 | 2 |
|---|---|---|---|
| Subject accuracy | Wrong subject or misleading scene. | Mostly correct with small ambiguity. | Clearly matches the requested concept. |
| Channel fit | Wrong crop or unreadable at target size. | Usable with editing. | Ready for the intended platform crop. |
| Rights safety | Uses risky names, logos, characters, or confidential assets. | Needs one safety edit. | Uses descriptive, rights-safe language only. |
| Need | How to tune the prompt |
|---|---|
| Product hero | Use material, surface, background, practical use case, and a clean negative prompt. |
| Social creative | Add audience emotion, hook space, crop ratio, and brand-safe visual constraints. |
| Consistent series | Lock palette, lens language, format, and recurring props across each variation. |
Score each result on subject accuracy, composition, channel fit, rights risk, text readability, and defect count. Regenerate only the weakest variable instead of rewriting the entire prompt when one dimension fails.
| Check | Pass condition | Fix if weak |
|---|---|---|
| Subject accuracy | The generated image matches the requested product, scene, or thumbnail promise. | Add material, shape, size, audience, and non-goals. |
| Rights and brand safety | No unauthorized logos, characters, living-artist imitation, or misleading claims. | Replace names with visual descriptors and remove risky claims. |
| Channel fit | The crop, focal point, and contrast work for the intended platform. | Name the aspect ratio, mobile crop, copy space, and focal hierarchy. |
| Defect control | No distorted text, warped faces, extra objects, or watermarks. | Tighten the negative prompt and simplify the composition. |
| Item | Snapshot |
|---|---|
| Page type | Existing Omellody AI image prompt utility refreshed in Red Mode; no new URL created. |
| Demand signal | Traffic radar on 2026-06-11 again ranked image prompt generator demand as the top GSC-gap signal, while V5 Red Mode required existing-page depth rescue instead of new URL expansion. |
| Originality | Omellody-created formula, browser-side builder, examples, review table, FAQ, and source snapshot. No external repository content copied. |
| Last reviewed | 2026-06-11 |
Source snapshot ID: 2026-06-11-2232-red-mode-image-prompt-depth-rescue.