ChatGPT
GeminiChatGPTGeminiRecreate a graceful Navratri portrait in ChatGPT or Gemini with a wine-red lehenga, ivory blouse, dandiya sticks, flowers and glowing Garba lights.
Quick workflow
Copy the prompt, customize any details, and generate in ChatGPT or Gemini.
Open ChatGPT or Gemini in your browser or app.
Upload your selfie to use your own face (or skip to make a new character).
Copy the prompt card above and paste it into the chat.
Press send to generate. Adjust any details you want.
Browse our curated library of tested AI image prompts or filter by your favorite tool.
Guide
This prompt generates a fashion portrait of a woman in a wine red Navratri outfit holding dandiya sticks, lit by natural window light. You get a clean, editorial-style image that captures the festive mood without looking like a cluttered stock photo.
AI models often push reds toward bright crimson or orange. To keep the outfit in that deep, moody wine red, the prompt uses descriptive anchors like "deep burgundy" and "dark maroon". If the fabric looks too bright in your first generation, add "desaturated red" or "dark wine red lehenga" to force the shadows to carry the color. Heavy fabrics absorb light, so mentioning "matte silk" or "velvet" also helps ground the shade and stops it from looking like shiny plastic.
Shooting near a window gives you soft, directional light. The prompt asks for diffused window light from the left to create a gentle shadow on the right side of the face. This adds depth to the heavy embroidery usually found on Navratri clothes. Flat lighting washes out the zari work and mirror details. Keeping the light source on one side shows the texture of the fabric and the metallic threads.
Hands holding thin wooden sticks are a known failure point. The model might generate three dandiyas, fuse the wood into the fingers, or bend the sticks at impossible angles.
Rest the props to save the hands
Ask the subject to rest the dandiya sticks over her shoulder or hold them loosely at her side. Gripping them tightly in front of the chest almost always results in merged fingers or extra sticks.
When you do want the sticks in hand, specify "holding two thin wooden dandiya sticks loosely in one hand". Giving the AI a specific number and a relaxed grip reduces the chance of it spawning a fistful of wood.
GPT IMAGE 2.5 for heavy textures
Use GPT IMAGE 2.5 for this specific look. It understands complex clothing textures like mirror work and heavy thread embroidery. Nano Banana Pro tends to smooth out fabric details, making the festive lehenga look like flat printed cotton.
GPT IMAGE 2.5 also respects the window lighting setup better, keeping the background naturally blurred while keeping the subject sharp. If you need to tweak the background, keep it simple. A blurred indoor room with warm fairy lights in the distance works well. Adding too many background elements distracts from the outfit and confuses the lighting engine.
FAQ
Tell the AI to place the subject near a large window with soft, diffused sunlight. If the shadows on the wine red lehenga look too harsh, add 'soft bounce light' or 'overcast window light' to your prompt. GPT IMAGE 2.5 handles natural light gradients well, so keep the lighting descriptions simple and direct.
Image models struggle with complex traditional patterns when you ask for too much at once. Instead of listing every type of thread and mirror work, just ask for 'fine zari embroidery' or 'heavy mirror work on wine red fabric'. Nano Banana Pro is good at keeping clothing textures sharp, but you might need to run the prompt a few times to get the folds and pleats looking natural.
Hands and small objects are still tricky for image generators. To get a clean shot, ask for the subject to hold the sticks loosely at waist level or rest them on a table nearby. If you specifically want them mid-dance, GPT IMAGE 2.5 usually gives better results when you describe the exact arm position, like 'right arm raised holding a dandiya stick'.
Next ideas
ChatGPTModel:GPT IMAGE 2.5
