How to Make a Gangnam Style AI Video from One Selfie

Selfie mode is designed for the joke of appearing beside yourself. The same uploaded face, hair and clothes guide all three copies. It does not assign three different outfits or let you direct each copy independently. The reference performance provides the group shots, camera movement and sequence of close-ups.

Choose this mode when you want one recognizable identity repeated across the scene. If you want yourself with two friends, switch to the trio mode and provide three portraits instead. Those photos have distinct Middle, Right and Left roles, explained in the photo selection guide.

Use a recent, well-lit photo with one person facing the camera. Keep both eyes and the mouth unobstructed, leave the hairstyle visible, and avoid sunglasses or a hand across the face. A portrait with some shoulders and clothing gives more context than a crop containing only facial features.

A mirror selfie can contain distractions such as a phone covering the face or a second reflected figure. Check the actual image rather than assuming every selfie is suitable. If the phone hides important details, take a simple camera-facing portrait instead. The upload does not need to have been taken with a front-facing camera.

Save it as JPEG, PNG or WebP at no more than 10 MB. Keep an unedited copy so you can try a different crop later without repeatedly compressing the same file.

1. Open the Gangnam Style studio and choose Three of you before adding your photo.
2. Upload the portrait and inspect its thumbnail. Confirm it is the right person and the face is clearly visible.
3. Select 480p or 720p. The current template uses fixed 16:9 landscape framing.
4. Review the quote against your current account or guest wallet balance and confirm permission to use the photo. Follow any sign-in prompt shown.
5. Submit once and follow the task status. Download the MP4 when generation and soundtrack preparation are complete.

Selecting another mode clears the current photo selection, so make the mode choice first. If you return to the page while a task is active, check the existing task before starting another. A new generation is a separate purchase using your account credits.

A 480p generation costs 150 credits; a 720p generation costs 300 credits. Selfie mode uses one upload, but the output is still the complete trio scene. It does not become a cheaper single-person clip because fewer photos are needed.

The video follows a 15-second reference and can finish slightly shorter. The reference clip’s soundtrack is added automatically, so you do not need to upload an audio file. Check playback with sound on once it is ready. For the practical difference between the two resolution settings, see the format and resolution guide.

Pause the opening group shot and check each copy, then watch the three solos and the final group. The intended identity stays the same throughout, but AI can vary facial features, clothing details or transitions. A strong first solo is not proof that every other shot matches equally well.

If the face is consistently hard to recognize, start by choosing a clearer portrait. If a task is still processing or needs support review, keep its generation ID rather than repeatedly submitting. The cost and fixes article covers that distinction and explains when failed-generation credits are returned.