Initially, I just wanted to use Gemini’s image generation AI to create a coherent comic, but I found the workflow to be incredibly tedious. To solve this, I built an automation tool using n8n to generate the panels. It worked by automatically cutting the storyboard and passing the previous image as context for the next panel's reference. However, this left me with very little control over individual shots—you couldn't "intervene" mid-process—which was quite frustrating. So, I spent over a week using a mix of manual coding and "Vibe Coding" to build an AI comic creation tool that allows for much more precise control. The editor is divided into three main areas: the Storyboard Zone, the Asset Zone, and the Generation Zone. Key Features: Asset Management: Users can upload their own resources, such as characters and backgrounds. Each asset can contain multiple images with descriptions that are sent to the AI as reference points. Seamless Context: You can pass a generated image from one panel directly to the next as a reference, eliminating the hassle of repetitive uploads. Precise Control: Unlike fully automated flows, this allows for manual intervention at any stage of the process. Currently, the tool is still just a Demo with plenty of bugs and unfinished features. If it gets a positive reception, I might continue to refine it and add more functionality.

ThirstyComic

Accurately generate comic by controlling each shot.

Accurately generate comic by controlling each shot.