A plain selfie can become a studio portrait, a retro snapshot, or a cinematic scene with a clear Gemini AI prompt for your own photo. The trick is to say what should change, what should stay, and what the final image is for.
We analyzed 77 comments and questions from YouTube, Reddit and Quora about Gemini AI photo prompts and found that 12% mentioned poor image quality and mismatched results.
Use the examples below as starting points, then adjust them to fit your face, setting, and goal.
Build a Clear Prompt Around Your Photo
A good prompt tells Gemini what’s in the photo and how you want it edited. Name the subject, the crop, the setting, the light, and the style. Then state which parts of your appearance should remain the same.
You don’t need every detail for every image. A headshot may need a crop and light direction, while a cinematic portrait needs a mood and setting.
Try this simple structure: “Edit my uploaded photo of [person] into [type of image]. Keep [identity details] unchanged. Set the scene at [place], with [lighting] and [style]. Use a [crop or framing].” Replace the brackets with plain, specific words.
For example: “Edit my uploaded selfie into a professional headshot. Keep my facial features, age, and skin tone recognizable. Use a dark blue jacket, soft window light, and a calm, natural expression. Crop from the shoulders up against a softly blurred office background.”
Be precise about identity. “Keep my face” is a start, but “keep my facial features, age, and skin tone” gives clearer boundaries. If you want a new outfit or background, say so directly. If you don’t want the pose to change, include that too.
Prompt writing follows the same basic idea in other AI tasks: clear instructions and useful context reduce guesswork. Coding Studio explains this in its guide to prompt engineering and clear AI instructions. For photos, the image itself is context, while your words set the edit.
Start with one main change. If you ask for a new outfit, a different pose, a new location, and a dramatic style at once, it can be harder to tell which instruction caused an unwanted result. Add extra details after you see the first version.
Adapt Prompt Examples for Different Photo Goals
Choose the prompt based on where you’ll use the image. A profile photo needs a clear face and calm light. A social post can carry a stronger mood, while a couple portrait needs instructions for both people.
Professional or LinkedIn portrait
“Use my uploaded photo to make a clean professional headshot. Keep my face, age, and skin tone recognizable. Dress me in a charcoal blazer and a simple light shirt. Use soft daylight, a neutral office background, and a relaxed expression. Crop from the chest up. Keep the result natural and suitable for a work profile.”
Keep the clothes close to what you’d wear at work. A dramatic fashion look may feel out of place on a company profile, even if the image looks polished.
Retro or cinematic edit
For a retro look: “Restyle my uploaded photo as a warm 1990s film portrait. Keep my face and pose recognizable. Add visible film grain, slightly faded colors, and soft direct flash. Keep the original setting.”
For a cinematic look: “Turn my uploaded portrait into a cinematic evening scene in Jakarta. Keep my face recognizable. Use warm light from a street-side café, a darker background, and shallow depth of field. Keep the expression natural.”
Aesthetic café portrait
“Edit my uploaded photo into a relaxed portrait at a café in Bandung. Keep my face and skin tone recognizable. Show me seated near a window with a cup on the table. Use soft afternoon light, warm colors, and a candid pose. Keep any signs or menus unreadable.”
Couple or pre-wedding portrait
Use a clear description for each person instead of saying “make us look romantic.” For example: “Create a pre-wedding portrait from our uploaded couple photo. Keep both people recognizable and keep their facial features and skin tones true to the references. Show the woman in a simple cream dress and the man in a dark suit, standing together in a garden in Yogyakarta at sunset. Use gentle warm light and a natural pose.”
For gender-specific wording, name each person as you intend: “the woman in the cream dress” and “the man in the dark suit.” Change those details to match your photo and preferences. If the result mixes faces or clothing, ask Gemini to correct only that detail.
Black-and-white or artistic portrait
“Make my uploaded portrait black and white. Keep my face and pose recognizable. Use soft side lighting, deep but detailed shadows, and a simple background. Add a fine art photography feel without changing my age or expression.”
Use this quick guide to match the image goal with its main visual instruction:
| Goal | Lighting to ask for | Framing or quality cue |
|---|---|---|
| Work profile | Soft, even daylight | Sharp head-and-shoulders crop |
| Retro snapshot | Direct flash | Film grain and slightly faded color |
| Cinematic portrait | Warm side light | Shallow depth of field |
| Couple portrait | Soft sunset light | Both faces clear in the frame |
| Black-and-white art | Side light with defined shadows | Fine detail in bright and dark areas |
These are prompt cues, not a promise that every output will have the same quality. A sharper source photo gives you more to inspect, but you should still check the generated face before using it.
For more reusable wording across work and personal tasks, Coding Studio has a collection of AI prompt examples for different needs. Take the structure, then rewrite the details for your own photo.
Upload Your Photo, Generate, and Refine the Result
The basic flow is simple: open Gemini, add a photo, describe the edit, and review the result. App layouts can change, so the name or position of an upload control may look different on your screen.
- Choose a clear source image. Pick a photo where your face is visible and not covered. Use an image you have the right to upload.
- Open the image attachment control. In the chat, look for the add or media option. Some layouts may place image choices inside a menu, then open your device’s media selector.
- Select the photo. Wait for the image preview to appear in the conversation before sending your prompt.
- Paste one prompt. Describe the edit and what should stay the same. Keep the first request focused on one goal.
- Send the request. Tap the arrow or send control, then wait for the result to finish loading.
- Check the image before saving it. Look closely at your face, hands, clothing, and background. If something changed in a way you didn’t ask for, request a targeted correction.
For example, if the background is wrong but your face looks right, don’t restart the whole prompt. Say: “Keep the person and clothing unchanged. Replace only the background with a quiet garden in Yogyakarta at sunset.” That tells Gemini which part to edit.
To change the mood, ask for one visible adjustment: “Make the light softer,” “switch to black and white,” or “use a warmer color tone.” Review the next result before adding another change. This makes it easier to see whether the edit moved in the right direction.
Use follow-up instructions to refine the background or style while asking Gemini to keep your appearance intact.
If the face doesn’t look like you, start with a clearer reference photo or restate the identity details. You can also remove extra style words and try again. Save the version that looks accurate, not simply the one with the most dramatic lighting.
Check Accuracy, Privacy, and Responsible Use
AI portraits can change small details even when the prompt asks to preserve them. Check the face first, then inspect clothing, accessories, fingers, and anything in the background that could mislead someone.
For a LinkedIn profile or work bio, compare the image with how you look now. Keep your real age and skin tone in the prompt, and avoid edits that suggest a job title, uniform, or setting you don’t have. A polished image should still represent you honestly.
For a couple or family photo, get permission before uploading another person’s picture. This matters even when you’re making a private draft. Don’t use someone else’s face to imply they attended an event or agreed to something when they didn’t.
Think about the photo before you upload it. Choose an image without visible ID cards, home addresses, private documents, or other details you wouldn’t want included in an AI chat. You can crop out those details first. Avoid sharing sensitive images when a simple, non-personal example can do the job.
When you replace a background, check the finished scene for signs, logos, or objects that make the image seem like a real event. If the portrait is heavily edited, don’t present it as an unedited record. This is especially important for workplace profiles and images used in business materials.
If you want to build prompting skills beyond photo edits, Coding Studio’s School of AI covers prompt and context work with tools such as Gemini and ChatGPT. Coding Studio is an AI Transformation House in Indonesia, helping professionals and companies apply AI through structured learning and mentoring.
Keep this check in mind: if the image changes how you look or suggests something untrue, revise it before sharing.
FAQ
How do I use my own photo in Gemini AI?
Open Gemini, attach a photo, then write what you want changed and what should stay the same. Choose a clear image where your face is easy to see. After Gemini generates the result, check your likeness and ask for a focused edit if needed, such as replacing only the background or changing the image to black and white.
What should I write in a Gemini prompt for a professional photo?
Ask for a head-and-shoulders crop, soft light, work-appropriate clothing, and a simple background. Also tell Gemini to keep your facial features, age, and skin tone recognizable. Name the look you need, such as a friendly company profile or a more formal CV portrait, instead of relying on “make it professional.”
How can I make a Gemini AI photo look more like me?
Start with a clear reference photo and tell Gemini which identity details to preserve. Keep the prompt focused, then review the face before changing other parts. If the result drifts, try a different source image or ask Gemini to correct the face while leaving the pose and background alone.
Can Gemini change the background of my photo?
Yes, you can ask Gemini to replace a background while keeping the person in the photo. Name the new setting and say which parts must remain unchanged. For example, ask for a Jakarta office background while keeping your face and clothes as they are. Check the result for odd edges or misleading details.
Is it safe to upload a personal photo to Gemini?
Use care with any personal image you upload to an online AI tool. Avoid photos that show private documents, addresses, or other sensitive details, and get permission before using another person’s image. Review the service’s current privacy settings and terms before sharing photos you consider sensitive.
Conclusion
Start with one clear photo and one specific edit, then inspect the face before you share the result. Try a professional or creative prompt above, adjust only what needs work, and keep the final image true to you.
