WEBVTT

00:00.163 --> 00:04.493
This is ComfyUI for Animation,
part two: Keyframe Exploration

00:06.723 --> 00:10.493
This workflow takes a character reference
and generates a key frame illustration

00:10.713 --> 00:14.743
at any camera angle you choose, with
an optional environment and action pose

00:16.813 --> 00:19.673
Load your character or character
sheet into the input node.

00:20.013 --> 00:21.963
Then adjust the multi-angle camera node.

00:22.293 --> 00:25.503
Drag the direction, elevation,
and distance controls to

00:25.503 --> 00:26.633
set the angle you want.

00:27.153 --> 00:29.703
The node generates a camera
description automatically from

00:29.703 --> 00:33.613
your position, which gets passed
into GPT Image 2 as the prompt.

00:34.343 --> 00:37.533
Optionally, load a background
environment into the second input

00:37.553 --> 00:40.793
to place the character in a scene
and describe a specific action

00:40.803 --> 00:42.163
in the additional action node.

00:42.503 --> 00:45.533
For example, a pose, an
expression, or something the

00:45.533 --> 00:46.853
character is doing in the shot.

00:47.403 --> 00:50.093
But feel free to turn either
option off if not needed.

00:51.393 --> 00:54.343
Run the workflow and the model
generates the key frame from

00:54.353 --> 00:55.703
all three inputs combined.