Image generation overview

Where to generate media in WJI and how the four generation modes differ.

WJI has one generation engine, reachable from several places. Wherever you open it, the form works the same way.

Where you can generate

Entry point	What it is
Idea Image Generation app	Inside a workspace. Results attach to the idea you have selected
Image Creator in the top nav	A standalone generator, not tied to a workspace
Image management dialog	A three-column dialog for working through a dataset’s media
Infinite Canvas	A different experience — references come from connections, not an upload list

The workspace app is the one to use for project work. It has three columns: the dataset sidebar on the left, the dataset contents in the middle, and the generator on the right. Anything you generate is filed against the current dataset automatically, and history defaults to filtering by it.

The Idea Image Generation app

The four modes

Icons at the top of the form switch between them.

Image

Still images from a prompt and optional references.

Video

Video, with duration, first and last frames, and task modes.

Storyboard

A grid of shots or camera angles in one image.

Text to audio

Speech from text, with speaker and rate controls.

Your mode choice and model selections are remembered between sessions.

The shape of a generation

1. Write a prompt

   See Writing prompts. WJI can rewrite it into an optimized final prompt before sending.

2. Pick a model

   The model determines which settings are even available. See Choosing a model.

3. Add references

   Optional, but this is how you keep characters and places consistent. See Reference images.

4. Check the cost, then submit

   The generate button shows an estimate before you commit. See How credits work.

5. Work with the result

   Results land in history below the form. See History and results.

A generation cannot be cancelled once submitted. There is no stop button — check your settings and the cost estimate before you press generate.