A Foundational Human Intelligence Mental Model for Visual Fluency.
SnapCharcoal doesn’t install new knowledge — it unlocks what’s already there by loading your entire visual library into your subconscious working memory context window.
It taps into the innate visual library etched into human perception: shapes, shadows, symmetry, and stories of line.
By re-wiring signal pathways, it allows the hand to bend the line to the will of the user, channeling unconscious visual memory into deliberate mark-making.
The reinforcement learning technique follows a deliberate path: it begins with basic drawing primitives — 3 lines, 3 shapes, 3 forms — to establish visual fluency. From there, it expands into broader visual concepts, layering in proportion, perspective, structure, light, and composition. Finally, it advances to the human figure, which offers a never-ending supply of training observations and the fastest eye–hand feedback loop for mastery. The result is a line drawn consciously while simultaneously considering all the visual possibilities it could take unconsciously. The resulting process is an instant human visual diffusion process that, for practical purposes, happens instantaneously.
It’s not magic. It’s memory, precision-tuned.
A specialized foundational model for visual fluency. Weights open-sourced. Runs only on human wetware.
Firmly installed intelligence on edge human wetware
SnapCharcoal