Fanon.io AI Architecting

Fanon.io AI Architecting

Role

Visual Designer

Visual Designer

Year

2023

2023

Category

Graphic Design

Graphic Design

Prototyping

Prototyping

At Standard Deviants, I helped solve a core challenge in AI-generated comics and animation: producing images at scale while maintaining consistency across characters, environments, and scenes.


The workflow used a language model to translate story beats into structured visual prompts, identifying details like shot type, composition, and character count. These prompts were passed to a visual engine, where layout templates established the base composition and an IP Adapter applied consistent environments and art direction using location-specific references.


To maintain character consistency, object detection and segmentation identified individual figures within each composition. These areas were then replaced using trained character LoRAs, allowing the system to generate varied scenes while preserving recognizable characters and environments across panels.

At Standard Deviants, I helped solve a core challenge in AI-generated comics and animation: producing images at scale while maintaining consistency across characters, environments, and scenes.


The workflow used a language model to translate story beats into structured visual prompts, identifying details like shot type, composition, and character count. These prompts were passed to a visual engine, where layout templates established the base composition and an IP Adapter applied consistent environments and art direction using location-specific references.


To maintain character consistency, object detection and segmentation identified individual figures within each composition. These areas were then replaced using trained character LoRAs, allowing the system to generate varied scenes while preserving recognizable characters and environments across panels.

The workflow began with a library of simplified panel templates that encoded composition, framing, and character positions. This spatial information helped maintain continuity across panels, such as keeping characters consistently positioned within a scene. The template was combined with a prompt from our LLM-based story engine to generate the initial base image

The workflow began with a library of simplified panel templates that encoded composition, framing, and character positions. This spatial information helped maintain continuity across panels, such as keeping characters consistently positioned within a scene. The template was combined with a prompt from our LLM-based story engine to generate the initial base image

Next, the pipeline introduced background and environmental information supplied by the story engine. This transformed the base image toward the intended setting and art style while preserving the composition and spatial relationships established by the original template.

Next, the pipeline introduced background and environmental information supplied by the story engine. This transformed the base image toward the intended setting and art style while preserving the composition and spatial relationships established by the original template.

Finally, segmentation identified each character region in the image, allowing our trained LoRA character models to be inserted into the appropriate spaces. Character assignments were inherited from the original template—for example, the character assigned to the blue channel would remain in that corresponding position throughout the pipeline.

Finally, segmentation identified each character region in the image, allowing our trained LoRA character models to be inserted into the appropriate spaces. Character assignments were inherited from the original template—for example, the character assigned to the blue channel would remain in that corresponding position throughout the pipeline.

The diagram illustrates the original structure of our interactive narrative prototype. Players completed a survey defining elements such as protagonist, setting, genre, art style, story length, and unpredictability. These inputs guided the language model in selecting a narrative structure and generating a branching, game-like story shaped by player choices.


As shown in the diagram, the original workflow required significant user input before the story could begin. I proposed replacing this with a single-prompt entry point that immediately generated an opening scenario. The system could then use early player decisions to progressively infer preferences, personalizing the narrative through interaction rather than upfront configuration.

The diagram illustrates the original structure of our interactive narrative prototype. Players completed a survey defining elements such as protagonist, setting, genre, art style, story length, and unpredictability. These inputs guided the language model in selecting a narrative structure and generating a branching, game-like story shaped by player choices.


As shown in the diagram, the original workflow required significant user input before the story could begin. I proposed replacing this with a single-prompt entry point that immediately generated an opening scenario. The system could then use early player decisions to progressively infer preferences, personalizing the narrative through interaction rather than upfront configuration.

Fanon.io Identity and Branding

Fanon.io Identity and Branding

I also developed the naming, branding, and visual identity for Fanon.io, including creating the Fanon name itself. I established the platform’s color palette and overall brand direction, then extended that identity into the product experience by designing interface concepts, platform layouts, and story navigation. This work helped define how users would discover, explore, and interact with stories while maintaining a cohesive identity across the platform.

I also developed the naming, branding, and visual identity for Fanon.io, including creating the Fanon name itself. I established the platform’s color palette and overall brand direction, then extended that identity into the product experience by designing interface concepts, platform layouts, and story navigation. This work helped define how users would discover, explore, and interact with stories while maintaining a cohesive identity across the platform.

Next Graphic Design Project

Next Graphic Design Project

Fanon.io Proprietary Storyworks: Character Design and Illustration

Fanon.io Proprietary Storyworks: Character Design and Illustration

Alongside my work developing Fanon.io’s generative storytelling systems, I created hand-drawn illustrations, character concepts, environmental designs, and visual world-building for several proprietary stories used for both platform testing and launch content. Projects included Sunblade, The Holy Order, and Iceblade, where I helped establish the visual identity of each world through character design, environments, story imagery, and key art illustrations. The work shown here represents a selection of the visual development and finished artwork created for these stories.

Alongside my work developing Fanon.io’s generative storytelling systems, I created hand-drawn illustrations, character concepts, environmental designs, and visual world-building for several proprietary stories used for both platform testing and launch content. Projects included Sunblade, The Holy Order, and Iceblade, where I helped establish the visual identity of each world through character design, environments, story imagery, and key art illustrations. The work shown here represents a selection of the visual development and finished artwork created for these stories.

Next Illustration Project

Next Illustration Project

© All Right Reserved, Jessica Lee

© All Right Reserved, Jessica Lee