Photo guide · One required input
Gangnam Style AI From One Photo
Start with one clear portrait for the center performer. Learn who fills the two side positions, what each extra photo changes, and what the finished 15-second video includes.
Yes, one photo is enough. Your portrait controls the center performer, while empty left and right positions use preset neutral adult companions. You can replace either companion with a separate authorized photo. The result is a 15-second, 16:9 MP4.
| Required | One main portrait |
|---|---|
| Optional | Up to two companion portraits |
| Frame | 15-second 16:9 MP4 |
01 · Understand the input
What does “from one photo” mean?
The main portrait supplies the required person reference. It gives the center role a face, hairstyle, clothing, and general appearance to work from. The left and right slots are optional. When either is empty, the tool fills that position with a preset neutral adult instead of copying the center person three times.
One main adult plus zero, one, or two separately uploaded adult companions.
Group opening, center close-up, right close-up, left close-up, then a center-led group finish.
The timing is fixed: the trio appears from 0–2 seconds, the center performs from 2–4.9 seconds, the right person from 4.9–8.4, the left person from 8.4–12.3, and the center returns for the final 2.7 seconds. This role order is why choosing the correct upload slot matters.
02 · Before you start
Choose a photo that stays useful in close framing
Use a normal portrait of one consenting adult, with the face turned toward the camera and lit evenly. Keep the eyes, nose, mouth, hairline, and jaw visible. A dramatic side profile can look good as a photograph but leaves more features for a video model to infer.
| Photos | What you control | What to remember |
|---|---|---|
| 1 | Center performer | Left and right use preset neutral adults. |
| 2 | Center plus left or right | The other side uses a preset neutral adult. |
| 3 | Center, left, and right performers | Each fixed role has its own portrait reference. |
Only upload yourself or adults who agreed to this AI use. Do not use children, public figures, strangers, or a photo you are not allowed to process.
03 · Follow the workflow
Set up the one-photo flow in five steps
These steps mirror the current form. There is no separate trio-style selector: the product uses one fixed three-person performance and lets you control people, background, and output quality.
- 01
Choose one clear main photo
Use a sharp photo of one consenting adult facing the camera. A chest-up portrait usually gives the model more useful face and clothing detail than a distant group shot.
One adultClear faceEven lightVisible shoulders - 02
Decide whether to add companions
The main photo is required. Companion 1 and Companion 2 are optional separate slots. Leave them empty when you only have one authorized photo; do not combine several people into the main-photo slot.
- 03
Check the left and right positions
If a side slot is empty, the tool uses a preset neutral adult for that position. Upload a separate authorized portrait only when you want to guide the identity in the left or right role.
- 04
Choose the background
Select Seoul, Convenience store, Office overtime, Beach wedding, or Hotel lobby. You can also describe your own setting in the custom background field; this flow does not accept a background image upload.
- 05
Review the setup and sign in
Confirm that each face is visible, everyone shown is an adult, and you have permission to use every upload. The tool asks you to sign in before submission and shows the credit cost before any generation starts.
04 · Read each control correctly
Know exactly what you get
The current product creates a 15-second landscape MP4. HD uses 720p, completed files are saved to My Creations, and the final MP4 includes the reference clip audio used for timing. It does not clone your voice or grant publishing rights to third-party music.
| Item | Current setting | What it means |
|---|---|---|
| People | Center required; left and right optional | Empty side slots use preset neutral adults. |
| Background | Seoul, four scene presets, or custom text | One continuous setting across the sequence. |
| Video | 15 seconds · 16:9 · 720p HD · MP4 | Download from My Creations after completion. |
| Price and account | Sign-in required; from $9.99 for one HD video | One-time credits; no subscription is required. |
See the current one-time packs on the pricing page. Credits are reserved when a generation starts and consumed only after success. Failed or blocked tasks release the reservation automatically; the refund policy explains quality-review and cash-refund rules.
05 · Judge the complete trio
Review more than the first clear frame
A useful review follows the main face from the opening group view through later close framing. Check whether hair, clothing, and distinctive features remain recognizable. Then look at the supporting performers, hands near faces, and moments when people overlap.
The site currently publishes a verified three-photo input-to-output example on the homepage. It was generated as a 14.869-second, 480p Beijing-night test with three separate portraits. It documents the sequence and slot order rather than the appearance of preset side performers in a one-photo result.
06 · Improve the input
Fix the photo before repeating the same setup
When a result looks weak, first check whether the source photo clearly contains the features you expected the model to preserve. A new attempt with the same ambiguous photo can repeat the same problem.
The face is too small
Replace a distant full-body or group image with a closer portrait. Cropping cannot recover detail that is missing from the original file.
The face is partly covered
Use a photo without sunglasses, a mask, hands across the face, or strong hair obstruction. Clear input reduces ambiguity but does not guarantee an exact likeness in every frame.
A companion does not match your upload
First confirm that the portrait was placed in the intended left or right slot. If the person is present but the likeness drifts, use a sharper front-facing image and read the face-matching guide before trying again.
The framing feels tight
The output is landscape 16:9. Choose a source photo with space around the head and shoulders rather than placing the face against an edge.
For identity drift, use the face-matching troubleshooting guide. If you want to control the left and right performers, read how to make a Gangnam Style AI video with friends. The exported MP4 already includes reference audio for rhythm, though it does not grant third-party music rights.
07 · Quick answers
Gangnam Style AI from one photo FAQ
Is one photo enough to start?
Yes. One clear adult portrait is enough. It controls the center performer, while empty left and right slots use preset neutral adult companions.
Who fills the other two positions?
When you upload only the center photo, the left and right positions use preset neutral adults. You can replace either one by adding a separate authorized adult portrait to that slot.
Will I appear in all three positions?
No. Your required photo guides the center role. The side roles use their own uploaded references or preset neutral adult companions when those slots are empty.
Can I add one friend instead of two?
Yes. Add one authorized adult photo to either available companion slot and leave the other optional slot empty.
Which photo works best?
Choose one sharp, front-facing adult portrait in even light, with the full face visible and no heavy filter, sunglasses, or mask.
What video do I receive?
The current product output is a 15-second landscape MP4 in 16:9. HD generation uses 720p and the completed file is saved to My Creations for download.
Does the video include sound?
Yes. The final MP4 includes the reference clip audio used for timing. The tool does not clone your voice or grant rights to third-party music, so confirm your right to publish the audio.
Is it free, and do I need an account?
Choosing photos and setting up the form is free. Generating a video uses credits, requires sign-in, and currently starts at $9.99 for one 720p generation.
What happens if generation fails?
Credits are reserved at submission and consumed after a successful completion. If a task fails or is blocked, the reserved credits are released to your balance automatically.
Is the downloaded video guaranteed to have no watermark?
This guide does not promise a watermark-free result. Check the completed MP4 before publishing it and contact support if the delivered file is materially different from the submission terms shown to you.
Is the video vertical?
No. Every video uses a landscape 16:9 frame.
Use only material you have permission to process and share. Read the terms and privacy policy before uploading another person.
Start with the photo you have
Choose one clear portrait.
Build the trio carefully.
Your portrait controls the center role; empty side positions use preset neutral adults. Add companions only when you have separate authorized photos, then review the fixed 16:9 setup before signing in and continuing.
Try the one-photo flow