Your selfie as a premiere photo: walking the red carpet between photographers or posing at the photocall under the flashes. Face kept, drift shown. Workflow visible here.
Real workflow: a man in a navy sweater against a grey wall, then the same man in a black tuxedo on a red carpet at night, photographers on the left and fans with phones on the right, and in a midnight blue velvet jacket in front of a burgundy backdrop with gold stars. On the carpet he smiles with his teeth showing and his shoes are out of frame. At the photocall, a photographer gets into the shot and a wedding ring appeared.
At a glance
The AI red carpet photo made from a selfie comes back several times a year: in May with gala season, during premiere season, then again before the holidays. You send your photo with a prompt and end up in a tuxedo under the flashes. The templates and guides we reread on 2 November 2026 give no audience figures, so there's nothing to back up talk of a huge wave. This workflow makes the two pictures that show up most: the arrival on the carpet between photographers and fans, and the solo pose in front of the photocall backdrop.
The model, Nano Banana Pro, keeps the demo's face, shave and haircut. It still takes liberties. On the carpet, the man smiles with his teeth showing when his photo has a closed-mouth smile, his hair comes out darker, and the frame stops at mid-shin instead of showing his shoes. At the photocall, a blurred photographer gets into the frame despite the prompt, and a wedding ring appeared on his hand.
For a cover with your own headlines, use the magazine cover. For a set of portraits without the evening theme, it's the AI photoshoot, and to try on formal wear for a ceremony, the wedding suit. If you'd rather act in a film scene than arrive at its premiere, see put yourself in a film. Both renders cost 38 credits, see pricing.
These are the files of the workflow above, exactly as the models produced them, unretouched. The starting point is shown when there is one.
Starting point
Render
People who want to do the AI red carpet photo trend for their stories or profile picture, friends planning a film-themed party, content creators following the formats that come around with galas and premieres.
A sharp, front-facing, well-lit selfie of you or of someone who agreed, and the exact colour of their hair written into the prompt. The photo is optional.
Take a sharp front-facing selfie
Face straight on, good light, no filter. A chest-up shot like the demo's is enough: the model makes up the body, the tuxedo and the setting.
Fix the hair description
The prompts say "short dark hair" when the demo's hair is brown, and the carpet render darkened it. Write the exact colour and length of yours, and add "closed-mouth smile" if that's how you smile.
Pick the carpet or the photocall
The carpet shows an arrival in motion, at night, with the crowd around. The photocall gives a posed, tighter portrait in the hard light of the flashes. Run both, or a single node for 19 credits.
Check the edges and the hands
In the demo, a photographer spills into the photocall frame and a wedding ring was added. If that bothers you, make the prompt stricter ("no ring", "nobody else in the frame") and rerun that node.
The same canvas, tuned for different needs: each variation is two or three lines of prompt away.
The carpet tells an arrival: he walks, the crowd surrounds him, flashes go off at the side of the frame. The photocall isolates the person against a plain backdrop, with a hard shadow and closer framing. The first makes a mood shot, the second a cleaner profile picture.
Change "the man" to "the woman" and the tuxedo to a long gown, with the colour, fabric and neckline spelled out. Add "no designer jewellery, no logo" to keep brands out. Not tested in the demo, and remember to update the face and hair description.
Start from a photo where both faces are sharp and rewrite the prompt with two guests, arm in arm, and one line of description each. Not tested here, and both people have to agree to their photo being used.
Replace "deep burgundy with gold stars" with your colours and a simple pattern, diamonds or circles. Keep "no letters, no logo": the model is quick to invent fake sponsor names that come out misspelled. Not tested in the demo.
Add at the end of the prompt "black and white photo, film grain, strong contrast, like a 1960s press photo". The red of the carpet goes, and the flash and shadow stand out more. This version hasn't been tested.
Connect a Video node to the carpet render and describe the motion: he takes a few steps, turns towards the photographers, the flashes pop. The demo stops at the two still images, this step hasn't been tested and the video costs extra credits.
The model has one photo to rebuild the face from. In profile, blurry or badly lit, it fills in the gaps, and under the hard flash light the difference shows even more.
In the demo, that line darkened brown hair. If you're blond, red-haired or have long hair, rewrite it, or the model blends your face with the description.
The image is meant to look like the person, and the setting suggests a real event. Only use photos of people who said yes, and never a famous person.
No, it's optional. Without one, the model invents a guest from the prompt. With your selfie, it keeps your face, which is the whole point of the trend.
You can. Here, both prompts are ready and wired to the same photo, the model is picked for face accuracy, and the demo shows you up front what it keeps and what it gets wrong.
The prompt asks for photographers out of frame, but the model still put one, blurred, on the right edge, with his lens and hand. Crop the image or rerun the node: the result changes from one try to the next.
38 credits, 19 per render on Nano Banana Pro. Rerunning a single image costs 19 credits.
Yes, by switching the node to Nano Banana 2.1 in 1K, 7 credits per render. The face is less faithful, especially at the photocall where it takes up more of the frame. That's why the demo runs on Nano Banana Pro.
1792 x 2400 pixels, vertical 3:4. That's enough for a story, a profile picture or a small print, not for a poster.
The prompt asked for the whole body, but the frame cuts the legs at mid-shin. Add "wide shot, feet and carpet visible below the shoes" and rerun if you want the full figure.
The starting photo has none and the prompt doesn't mention one: the model added it on its own. Write "no ring, no jewellery" to avoid it.
Make an AI red carpet photo from your selfie
Use this workflowFree account, no card required. The workflow opens pre-filled in your canvas.
Nano Banana Pro
Flux Kontext
Seedream 5 Pro