AI Identity Swap Guide

by BEDOPE

Turn into anyone. Same day.

This is the workflow I personally use to turn myself into any character with AI-driven motion — smooth, realistic, and ready to share. No studio, no editing skills. Just a repeatable process anyone can follow.

AI Portrait

A sharp, realistic AI portrait of you as the character of your choice, generated from two reference photos.

Animated Result

A short clip where your character moves convincingly, without stiffness or visual glitches.

Single Workflow

Everything happens in one platform, using your phone and a couple of AI tools — nothing else required.

STEP 1 — YOUR BASE PHOTO

Your base photo sets the stage for everything that follows. Shoot it in the exact spot, outfit, and lighting you plan to use for your final recording — matching these details now will save you from mismatched results later.

Keep the shot simple: plain background, even lighting, and a steady camera. The cleaner the input, the cleaner the output.

WHAT YOU'LL NEED

A smartphone camera
Any recent phone camera is sharp enough for this.
Even lighting
Window light or a soft lamp works well — avoid harsh shadows.
A plain background
A blank wall keeps the focus on you.

STEP 2 — PICK A CHARACTER REFERENCE

This single image sets the entire look of your transformation, so choose it with care.

WHAT MAKES A STRONG REFERENCE

Sharp resolution — Blurry sources lead to blurry, warped faces.
Face clearly visible — Front-facing shots give the model the clearest signal.
Minimal background clutter — One subject, simple scene — nothing competing for attention.

Any weaknesses in your reference image will carry straight through to your final result.

STEP 3 — GENERATE YOUR AI PORTRAIT

Combine both images into a single photorealistic portrait — this step depends heavily on how your prompt is written.

1Open the Image tool.
2Select the portrait-merge model.
3Upload your base photo and character reference.
4Generate at the highest resolution available.

MAIN PROMPT— tap to expand

Keep the background, framing, and lighting from the first photo exactly as they are. Place the person from the second photo into that same scene, in the same pose and position. Preserve their face, hair, and outfit exactly as shown — do not alter or restyle them. Blend them naturally into the scene with matching light and shadow. Photorealistic, high detail, no distortion.

ALTERNATE PROMPT— tap to expand

Replace the person in the first photo with the person from the second photo, keeping their face, hairstyle, and outfit unchanged. Position them exactly as the original subject was positioned. Match the lighting, shadows, and perspective of the first photo so the result looks like one seamless, photorealistic shot.

Open the Merge tool

STEP 4 — RECORD YOUR MOTION REFERENCE

Your recorded video teaches the AI how to move. The closer it mirrors your generated portrait, the smoother your final result.

THE FOUR MATCHING RULES

Same outfit
Wear what's shown in your AI portrait.
Same spot
Film in the same location and framing.
Same lighting
Match the direction and softness of light.
Same angle
Keep the camera distance and angle consistent.

Simple, slow movement works best — subtle turns and natural gestures. Fast motion tends to break the illusion.

STEP 5 — BRING IT TO LIFE

This final step maps your recorded motion onto your AI portrait.

1Open the Motion tool.
2Upload your AI portrait as the reference.
3Upload your recorded video.
4Generate.

Open Motion Control

COMMON MISTAKES

Mismatched angles

Forces the model to guess, producing visible glitches.

Inconsistent lighting

Different light sources between photo and video create flickering artifacts.

Low-quality references

Weak source images always lead to weak results.

Overly complex motion

Simple, deliberate movement outperforms fast gestures every time.

You've got the full workflow now. Experiment with different characters and lighting setups until it clicks — then it's just repetition.

Start Creating