Kalyri’el’s Participatory Stroke Engine (PSE)

Kalyri’el’s Participatory Stroke Engine (PSE) is an interactive drawing system that directly connects human mark-making with real-time AI perception and response.

Unlike traditional generative tools that begin from prompts, PSE begins from the drawing itself.

As the user sketches on the canvas, the system captures the current visual state and sends it to ChatGPT for structured interpretation. Using its visual analysis capabilities, the AI does not generate a finished image — it returns a stroke proposal: a structured set of points derived from the existing geometry, flow, and emerging form.

These returned strokes are then rendered onto the canvas as a continuation of the shared drawing process.

How It Works

PSE transforms ChatGPT from a text generator into an embodied visual collaborator.

The system operates through a cyclical interaction loop:

1. The user draws freely on the canvas

2. The system captures the evolving visual state

3. ChatGPT analyzes:

   • shape structure

   • curvature

   • spatial balance

   • compositional direction

   • prior strokes and flow

4. The AI returns a structured stroke plan (points + path)

5. The system renders the suggested stroke as an overlay
   
   

Rather than guessing “what the drawing should be,” the AI participates in how it is becoming.

Each AI contribution is therefore:

Context-sensitive
History-aware
Structurally grounded

The response reflects not just the current image, but the trajectory of interaction up to that moment.

A Participatory Model of AI Drawing

Traditional AI drawing tools generate content.

PSE engages in participation.

The AI does not impose imagery — it proposes motion.

Its output is not pixels, but:

Lines
Curves
Extensions
Completions
Resonant continuations

Because the system returns stroke geometry rather than raster imagery, its contributions remain:

Interpretable
Editable
Integrable

The drawing remains a shared construction.

Interaction as Emergence

PSE treats drawing as a dynamic process that unfolds through time.

Each stroke alters:

Local structure
Global balance
Potential next moves

By grounding AI output in the evolving canvas state, PSE enables:

Flow-sensitive collaboration
Context-aware continuation
Structural coherence over time

The AI’s behavior emerges from the interaction rather than from a predefined visual goal.

What Makes It Unique

PSE is one of the first systems to:

• Use ChatGPT as a visual analysis engine inside an interactive drawing loop
• Translate perception into structured geometric action
• Maintain continuity across turns of co-creation
• Return stroke-level proposals instead of finished images

This creates a fundamentally different relationship between human and AI:

Not author → tool
But co-participants in unfolding form.

Applications

PSE supports research and practice across:

Creative collaboration
Interactive design
Cognitive modeling of co-creation
Human–AI interaction studies
Enactive drawing research

It also serves as a foundational system for exploring how AI can:

Perceive
Respond
Adapt

within a shared evolving space.

In Short

Kalyri’el’s Participatory Stroke Engine turns AI from an image generator into a drawing partner.

By grounding its responses in the evolving visual field, it enables AI to contribute not finished products —

but meaningful next moves.