A storytelling tool kept turning into a systems problem.
MiMEngine started several years ago as a way to build comics and tell an original story. As the project grew, the problem expanded from page layout into characters, world state, lore, locations, events, timelines, media, publishing, reader experience, and eventually AI.
Rather than reduce that complexity to a single prompt box, I kept building the underlying system.
Phone-first sequential storytelling, not a PDF scroller.
The reader experience is designed around sequential art on a phone: swipes reveal pages and panels can move through text-to-action, action-to-action, and moment-to-moment transitions rather than simply shrinking a printed comic onto a small screen.
On the publishing side, Drupal and Drupal Commerce provide a path for creators to publish stories, manage access, and support subscriptions or purchases.
Make the orchestration visible enough to understand.
The current system explores specialized agents and tool servers that can work on different parts of an author's world. A request might concern a character, location, event, lore item, timeline, or story beat. The workflow passes structured context between components and can produce machine-readable output such as JSON.
A key design goal is transparency: show the log, the handoff, the tool or agent involved, and why a decision was made. Creative AI becomes more useful when the author can inspect the process rather than simply accept a mysterious result.
Preserve author intent while increasing creative leverage.
MiMEngine is not an attempt to automate the author out of authorship. The harder problem is deciding where models should propose, where tools should transform, where structured rules should constrain, and where the human should intervene.
The project uses generative image/media tooling, inpainting/refinement ideas, multiple frontier models, local/open-source components, and agent orchestration as parts of a larger creative workflow.
An end-to-end system and an ongoing R&D engine.
MiMEngine has also become a credibility engine for my broader work: museum animation projects and generative-media experiments grew out of the same creative-technology exploration.
The project is an opus, but the career lesson is practical.
- Learn emerging technology by forcing it into a real product.
- Design systems around creative intent, not just model capability.
- Orchestrate specialized tools and agents instead of assuming one model should do everything.
- Think beyond prototype. Authoring, media, reader experience, publishing, commerce, and operations all have to connect.