How to Use Gemini AI Photo Prompts With Your Own Photo

Most prompts on PromptAI are written for one simple workflow: you upload a photo of yourself to Google Gemini, paste a prompt, and Gemini rebuilds the scene around your face. This guide walks through that process from start to finish, including what to do when the result does not look like you.

What you need

• The Gemini app on your phone, or gemini.google.com in a browser, signed in with a Google account.
• One clear photo of yourself (or of both people, if you are using a couple prompt).
• A prompt from the site. If you are not sure where to start, the portrait prompts and professional prompts are the most predictable.

The same prompts also work in ChatGPT's image generation, and the steps are almost identical.

Step 1: Choose the right photo

The photo you upload matters more than any word in the prompt. Gemini can only keep features it can see clearly.

1. Use a photo where your face is sharp, in focus and facing the camera or turned slightly.
2. Pick even, natural light. Harsh shadows across half the face often lead to a different-looking person.
3. Remove sunglasses, masks and hands in front of the face.
4. Avoid beauty filters and heavy edits. The AI will copy the filtered skin and it will look artificial.
5. For couple prompts, use one photo with both of you visible at a similar size, rather than two separate selfies.

A recent photo is best, especially for headshots you plan to use on LinkedIn or a resume.

Step 2: Upload the photo and paste the prompt

1. Open a new chat in Gemini. Starting fresh avoids old instructions affecting the result.
2. Tap the plus or attachment icon and upload your photo.
3. On PromptAI, open a prompt such as Corporate Executive Look for Man and copy the full text.
4. Paste it into the message box with the photo attached and send.

Gemini usually returns an image within a few seconds to a minute, depending on demand.

Step 3: Adjust the prompt before you send it

You do not have to use a prompt exactly as written. Small edits make the result feel like yours:

• Outfit: change colours or garments. "Navy blue kurta" can become "maroon kurta" with no other changes.
• Location: swap the setting. In Beach Surfboard Portrait, you could replace the coastal village with Goa or Varkala.
• Details you don't have: if a prompt mentions a beard, glasses or long hair and you don't have them, delete those words. Otherwise Gemini may add them.

Change one or two things at a time so you can see what each edit does.

Step 4: Keep your face consistent

Most prompts on the site already include a line like "keep the face 100% the same as the uploaded photo". Leave that line in. If you write your own prompt, add a similar sentence near the start.

If the face still changes, try these in order:

1. Reply in the same chat: "Keep my face exactly as in the original photo. Only change the outfit and background."
2. Upload a clearer, more front-facing photo and try again.
3. Shorten the prompt. Very long prompts with many details sometimes push the face lower in priority.
4. Start a new chat. After many rounds of edits, results can drift further from your original face.

Step 5: Fix common problems

AI image tools still make predictable mistakes. Here is how to handle the usual ones with a short follow-up message.

• Extra or strange fingers: "Fix the hands so each has five natural fingers." Cropping the final image tighter also helps.
• Wrong outfit: repeat the outfit clearly. "He should be wearing a cream sherwani, not a suit."
• Garbled text: for prompts with captions or jersey names, keep the text short and ask Gemini to "correct the text to read INDIA".
• Face looks pasted on: ask for "matching light on the face, same warm tone as the background".
• Too dark or too bright: "Add a soft fill light on the face" or "reduce the brightness of the background".

Fix one problem per message. Asking for five changes at once tends to create new problems.

Step 6: Iterate, then download

Treat the first result as a draft. Generate two or three versions, keep the best, and refine it with small requests. A Casual Studio Portrait might need only a tweak to the pose, while a busy scene like an Autumn Picnic Couple Portrait may need a couple of rounds to get both faces right.

When you are happy, open the image and use the download option to save it to your device. If you plan to print it, check the resolution first, as generated images are sized for screens.

FAQ

Is Gemini image editing free?
Gemini offers image generation and editing to free users with usage limits. Google changes these limits from time to time, so check the app for current details.

Why does my face look different in the result?
Usually because of the source photo: low light, an angled face, sunglasses or filters. Use a sharp, front-facing photo and keep the face-preservation line in the prompt.

Can I use these prompts in ChatGPT?
Yes. Upload your photo in ChatGPT, paste the same prompt and follow the same steps. Results will look a little different because the models are different.

Can I try a different hairstyle or outfit with the same prompt?
Yes. Edit the outfit or hair description before sending, or ask for the change in a follow-up message. The Hairstyle Grid Collage prompt is a quick way to preview several hairstyles at once.