Use a shot recipe
- Subject: age-presenting adult, body type, hair, face, outfit, and signature details.
- Camera: portrait, full-body, close-up, candid, mirror selfie, editorial, cinematic still.
- Setting: bedroom, gym, city street, hotel room, studio, beach, office.
- Light and mood: warm lamp light, neon, daylight, flash photo, soft shadows, high contrast.
- Composition: side view, centered portrait, over-the-shoulder, waist-up, wide shot, blurred background.
Keep a character consistent
- Repeat the most important identity traits in every image prompt.
- Use one stable outfit or accessory if the generator drifts.
- Avoid changing too many variables in one request.
- Save prompts that worked and make one variation at a time: setting first, then outfit, then lighting.
Borrow the Candy and OurDream image tricks
- If a prompt is ignored, emphasize the critical phrase with parentheses and move it earlier in the prompt.
- If the image is bland, add camera angle, lighting, composition, and mood before adding more adjectives.
- For video-style prompts, keep it to one focused scene, one main action, and one pacing direction.
- Small styling details such as makeup, accessories, or a signature color can anchor a character visually.
Copy-ready prompts
Create an image of my adult AI companion: [identity]. Shot type: [portrait/full body/candid]. Composition: [centered/side view/close-up]. Setting: [place]. Lighting: [light]. Mood: [mood]. Keep the face, hair, and body type consistent.
Revise this image prompt for consistency. Preserve the companion identity, simplify conflicting details, and make the camera direction clearer: [prompt].
Give me five image prompt variants for the same companion. Keep identity unchanged. Change only one variable per variant: setting, lighting, outfit, camera, or mood.