What kind of photo works best for Create Face?

A clear, well-lit, front-facing photo in which the whole face is visible. Avoid sunglasses, heavy makeup, extreme angles and anything that covers the face, because they reduce how well the identity is kept.

What is the difference between Create Face and Character Asset?

Create Face holds only a facial identity, made from one photo. Character Asset holds a whole character, with a portrait, expressions, poses, angles, motion clips and a voice.

Can one Create Face node hold two people?

No. Use one node per person, so each face keeps its own identity.

Create Face node — a reusable face identity from one photo, for face swap and consistency

Q: Where can I use the face?

Connect the node's Face output to the face input of Face Swap, or to the Assets input of Generate Image to keep the same face in new images.

Make a reusable face identity from one clear photo, and use it to keep the same face in generated images and to swap it into video with Face Swap.

The Create Face node makes a reusable facial identity from one clear photo of a person. Use the face to keep the same person in generated images, or to put the face on someone in a video with Face Swap. Unlike the Character Asset node, it holds only the face, not a whole character.

When to use it

Swap a real person's face into a video with Face Swap.
Keep the same face across several generated images.
Make a reusable face for talking-head content.
Keep a speaker's identity when you make video from their narration.

Quick start

Add the node

Press Tab on the canvas and choose Assets › Characters › Create Face.

Name the face and add the photo

In the settings panel, type the Face Name. Under Reference Photo, upload a clear, front-facing photo, or paste its URL.

Generate the headshot

Choose the Style and the Image Model, and click Generate Headshot. The button shows its price in credits. The headshot appears on the node.

Use the face

Connect the node's Face output to the face input of Face Swap, or to the Assets input of Generate Image.

A face made from one photo is swapped into an uploaded video.

Inputs and outputs

The node has one input, Prompt, on its left edge. It accepts text nodes and pickers.

The output, Face, carries the face identity. Connect it to the face input of Face Swap, or to the Assets input of Generate Image. When the face reaches an image model, its default role is face, so the model takes the facial features from it. You can change the role on a mention's pill. See Reference roles.

Settings

Setting	What it does
Face Name	The name of the face. The node needs a name to run. If another Create Face node in the workflow has the same name, a number is added, such as "John Smith (2)".
Description	Optional notes about the face or the context, such as "a person in their 30s with brown eyes".
Style	The visual style of the headshot: realistic (default), anime, 3d-pixar or illustration.
Reference Photo	The photo of the face. Upload it, or paste its URL. The node needs it to run.
Image Model	The image model that makes the headshot.

Type {} in the Face Name or Description field to insert the text of the node connected to the input.

Choose a good photo

Front-facing and well lit. The whole face is visible and evenly lit.
Nothing in the way. No sunglasses, hands or hair over the face.
No extremes. Avoid heavy makeup and extreme angles, which reduce identity accuracy.
Mouth visible, neutral. For lip-sync work, use a photo where the mouth is clearly visible and relaxed.

Tips

One face per node. Make a separate Create Face node for each person.
Generate before you connect. Run the node first, so the nodes that use the face have a finished headshot.
Reuse it. A face is saved to the project and can be used in other workflows of the same project.
Need more than a face? For expressions, poses, angles and a voice, use the Character Asset node.

Create Face