Comics creation and illustration with OpenAI's LLMs and GPT Image 1
This workflow relies fully on OpenAI's LLMs and vision models to generate the story line, the dialogues and illustrations of a comics about data analytics superheroes with incredible KNIME skills.
The workflow prompts a reasoning model to get a story outline, then it uses it to prompt a chat model to get a detailed story line with details and descriptions for each scene. At the same time, the vision model takes care of illustrating the characters.
A third call to an LLM generates the dialogues of each scene and a new call to the vision model illustrates those scenes, taking as input the previously generated illustrations of the characters for consistency.
Finally, all the scenes (dialogues and illustrations) are compiled into a static PDF report.