GPT Image 2 powers graphic novel workflow
The post shows a creator using GPT-Image 2 in multiple passes: first to convert prose into a graphic novel script, then to generate an initial image, and finally to start a new image session with the previous output as a style reference. The point is not perfection, but demonstrating that the model can support a chained creative workflow with continuity across sessions.
Hot take: this is a better demo of workflow design than raw image quality. The interesting part is the handoff between reasoning, generation, and style anchoring, which is closer to how real creative production works.
- –Shows a practical multi-step pipeline, not a one-shot prompt.
- –The reference-image follow-up suggests style consistency is becoming a first-class capability.
- –Useful for storyboarding, comics, and iterative visual development where continuity matters.
- –The Reddit framing makes it feel more like a creator workflow experiment than a polished product launch.
DISCOVERED
45d ago
2026-04-27
PUBLISHED
45d ago
2026-04-27
RELEVANCE
AUTHOR
semibaron
