Skip to main content

Creating with Performances on invideo

Updated yesterday

Introduction

Performances is invideo's AI-powered face and scene swap tool, built for creators and filmmakers who want to place themselves or others into any scene, swap characters, and create studio-grade content without a studio.

How to get started

  1. From the homepage, click Trends β†’ See all β†’ select Performances

  2. Upload your source face (the face you want to swap in) and your target image (where the new face will appear)

  3. Write the prompt "swap face" and hit generate

  4. Download your result instantly - no watermark, up to 4K

What you can do with it

Scene swap

Place any character into a completely different location, including abstract or AI-generated environments, in a single click.

Cast swap

Accurately interchange characters within a video without affecting the rest of the scene.

Head swap

Swap entire heads between images for transformations, VFX pre-vis, or creative effects.

Gender swap

Generate a hyper-realistic male or female version of any face.

Expression transfer

Preserves the original facial expressions from the source and transfers them seamlessly to the swapped face.

AI-generated backgrounds

Transform basic home footage into high-quality epic stage environments.

πŸ’‘ Tips for better results

Why it helps

Use a clear, well-lit front-facing photo as your source face

Gives the model the best reference for a realistic swap

Use Nano Banana Pro as your image model

Produces the most accurate skin texture, lighting, and blending

Keep your target image high resolution

Higher input quality = higher output quality

Write a specific prompt beyond just "swap face"

Describing lighting, mood, or style helps the model match the scene more precisely

Ethics and privacy

Invideo treats this seriously. Your images are never saved or used to train AI models. Every AI actor's likeness is used only with verified consent, and face-matching technology blocks any unauthorised use. Content moderation runs on every generation, human and AI combined.

Did this answer your question?