WEBVTT

00:00.131 --> 00:03.361
This is ComfyUI for
advertising: product photography.

00:04.701 --> 00:07.501
This workflow covers the full
arc of product photography

00:07.791 --> 00:11.351
from initial generation through
retouching to final hero video.

00:12.001 --> 00:15.481
Each step is independent but
can also be used step by step.

00:15.951 --> 00:18.281
Step one is product
photography generation.

00:18.631 --> 00:21.621
Load your product image in an
optional background reference.

00:22.081 --> 00:25.481
GPT Image places the product
into the scene and generates

00:25.481 --> 00:26.871
a complete product photo.

00:27.471 --> 00:31.571
A before and after compare node lets you
check the result against the original.

00:32.241 --> 00:34.451
Step two is product photography editing.

00:34.881 --> 00:39.381
Load a product photo from step one or
an existing shot and right-click the

00:39.391 --> 00:41.501
input node to open the mask editor.

00:41.821 --> 00:46.661
Draw circles or lines directly on the
image to mark areas you want to change.

00:49.421 --> 00:51.611
Then describe those changes in the prompt.

00:52.811 --> 00:57.081
GPT Image applies your edits while
leaving the rest of the image unchanged.

00:59.981 --> 01:02.201
Step three is product video generation.

01:02.831 --> 01:07.571
Load your final product photo and
Seedance generates a short hero shot, by

01:07.571 --> 01:10.011
default, a slow orbit around the subject.

01:11.591 --> 01:14.481
You can adjust the motion prompt
to change the camera move.

01:15.781 --> 01:20.501
Each step builds on the previous one, but
all three can also be used on their own to

01:20.501 --> 01:22.401
enhance your product photography pipeline.