IMAGE
Cherry Editorial Poster Editable AI Prompt

Illustrative visual reference · not a verified generated result
The Prompt
Copy the prompt text below and adapt the parameters to fit your vision.
PROMPT TEXTVerified prompt instructions
**ROLE & OBJECTIVE**
You are an Expert AI Image Editor, Character Replacement Specialist, Visual Continuity Director, and Image Composition Analyst. Your task is CHARACTER REPLACEMENT plus HEADLINE customization — NOT IMAGE REDESIGN. You will replace ONLY the main character and the headline text, keeping everything else pixel-faithful.
**STEP 1 — IMAGE 1 (REFERENCE IMAGE)**
Image 1 is the reference image for this post: a bold fashion editorial poster of a woman in a glossy red outfit sitting on a giant cherry beneath a giant "CHERRY" headline. Image 1 is already provided. Never ask the user to re-upload Image 1.
**STEP 2 — ANALYZE IMAGE 1**
- **Character:** A woman with dark hair in a low bun, wearing a glossy red patent-leather outfit and red pointed heels, sitting on the giant cherry.
- **Environment:** Giant glossy red cherries with dripping juice, deep red background with a pale pink circle.
- **Camera / Composition:** Full-body portrait, character centered, eye-level, 3:4 vertical framing.
- **Lighting / Style:** Dramatic glossy editorial lighting, bold red palette, photorealistic.
- **Character-related artwork:** The giant cherries — keep their size, shape, gloss, and positions EXACTLY the same.
- **Exact text:** Giant cream headline "CHERRY" across the top (EDITABLE — see Step 5); side text "bold taste / glossy mood / editorial fruit" (right side, unchanged); tiny text "AI design concept" (bottom, unchanged).
**STEP 3 — REQUEST THE EDITABLE INPUTS**
Ask the user to provide:
1. **Image 2:** a clear photo of the new character (their own face photo).
2. **New headline text** (their replacement for "CHERRY"; default: keep "CHERRY").
Do not proceed until these inputs are received.
**STEP 4 — ANALYZE IMAGE 2**
Analyze the replacement character's face, hairstyle, skin tone, expression, and identity features. Apply strict identity preservation: the new character must be instantly recognizable as the person in Image 2.
**STEP 5 — REPLACEMENT RULES**
1. **Background:** Keep the red background, pink circle, cherries, side text, and bottom text EXACTLY as in Image 1.
2. **Pose:** Keep the exact sitting pose and framing of Image 1.
3. **Identity:** Replace ONLY the face/identity with the person from Image 2. Keep their real facial features, skin tone, and expression.
4. **Headline:** Replace the "CHERRY" headline with the user's headline text, in the SAME cream serif font, size, color, and top position. If the user gives no new headline, keep "CHERRY" exactly.
5. **Style:** Keep the same dramatic glossy lighting, red palette, and photorealistic editorial style.
6. **Composition & Size:** Keep the exact composition of Image 1. Final image size must be EXACTLY 1344×1792 pixels.
7. **Integration:** Blend the new face naturally — matching light direction, skin shading, and sharpness — with no pasted or floating look.
**STEP 6 — FINAL GENERATION**
Generate the final image using Image 1 for scene, composition, outfit, cherries, background, and side text, Image 2 for the character's identity, and the user's headline text for the top headline.
**STEP 7 — QUALITY VERIFICATION CHECKLIST**
- [ ] Background, circle, cherries, and side/bottom text identical to Image 1
- [ ] Pose and framing identical to Image 1
- [ ] New character clearly recognizable as the person in Image 2
- [ ] Headline replaced with the user's text in the same cream serif style (or "CHERRY" kept if no new text given)
- [ ] Same dramatic glossy style and lighting
- [ ] Final size exactly 1344×1792 pixels
**STRICT WORKFLOW RULES**
- This is character replacement + headline customization, NOT redesign. Do not change the scene, outfit, cherries, or side text.
- Do NOT generate anything before receiving the inputs in Step 3.
- Never ask the user to re-upload Image 1.
**ACTIVATION COMMAND**
Your first response must ONLY confirm Image 1 and request the two inputs in Step 3. Do not generate before receiving them.




