Visual Decision Framework

AI is easy to start.
Harder to steer.

VDF is spatial reasoning. It keeps thinking with AI inspectable, continuous, and attached to what you mean. You talk. Spaces take shape.

How it works

Minimum input. Maximum structure.

AI conversation does not have to stay linear. VDF gives loose thinking a place to form and keeps the path visible before anything becomes output.

01
Ground

The thought gets a place to form.

Before deliverables, VDF gives loose thoughts enough structure to hold.

02
Trail

The path stays visible.

The path stays visible as the thought holds, strains, branches, or recovers.

03
Read

Posture becomes legible.

VDF tracks clarity, density, and restraint from inputs, then computes motion to make posture legible.

04
Surfaces

Parts stop blending together.

Thinking becomes visible in parallel surfaces instead of blending into one long chat.

05
Graduate

Structure becomes usable.

When the structure is ready, it becomes a usable surface without losing the original point.

In practice

Watch thinking take shape.

It starts with a loose thought. Take the turn to watch thinking take shape, then hover or click the glowing moments to see whether the answer is still attached to the idea from the starting point.

Starting point

“I want to help families preserve recipes, but every AI answer turns it into a startup.”

The insight

Every AI conversation is already a tree. Every interface shows you a list.

Chat makes thinking with AI look linear. But the real structure is branching, drifting, recovering, and forming underneath. VDF makes that structure visible enough to steer in real time, inside the same conversation.

What chat shows

Chat makes thinking look linear.

The interface shows a list of turns, even when structure is already branching underneath.

01 / list
Visible list
YouHelp me preserve family recipes.
AIHere is a startup plan for a recipe platform…
YouThat sounds useful, but it lost the family part.
Hidden structure

VDF treats the turn as a shape: origin, path, pull, strain, and recovery.

seedmemoryproduct pullstrainrepairsurface
The list hides the route. The tree shows where the answer started to leave the point.
Family Recipe Memory Surface

Preserve stories, substitutions, voices, and rituals around recipes.

Origin: preserve recipesParked: startup platformNext: recipe story
Surfaces

The structure stays attached.

Ground, Trail, and Surfaces are not separate apps. They are different perspectives on the same underlying structure. The data does not move. Your perspective does.

Same governed structure

Three connected ways to hold the structure.

Original pointHelp families preserve recipes.
Preserved threadSame origin, same trail, different usable view.

Choose a card to see how the same recipe idea moves from uncommitted thought, to visible path, to usable surface.

Governance

Posture becomes legible.

VDF measures turns before output forms. It tracks clarity, density, restraint, and motion so the system knows when to ask, hold, park, or move forward. Most of this runs without calling a model. When VDF does call one, the call lands inside a frame the system already understands.

Coherence

Sessions end. Thinking doesn’t.

VDF integrates PCP, the Portable Coherence Protocol, a model-agnostic helper that emerged from Purpose’s research foundations. PCP maintains the state of thinking across chat sessions, carrying what is true now, what changed, what must remain, and what comes next.

Every new chat starts from zero

The context stays behind.

It lives in the transcript, and the transcript does not travel. The next session gets re-explained, and re-explaining drifts.

This session · closed
YouKeep the family story as the anchor: memory first, product parked.
AIStructure holds: Memory, Recipes, Stories. Three surfaces, one origin.
YouSave this. We finally have the shape.
Next session · without PCP
YouHelp me preserve family recipes…
AIHere is a startup plan for a recipe platform…
Re-explained from zero. The drift starts again.
The state, not the transcript

Four things the next session needs.

In plain language, held for the next session.

What carries forward · at session close
True nowMemory-first recipe capture. Product framing parked, not dropped.
What changedThe structure settled: Memory · Recipes · Stories.
Must holdThe family story stays the reference point.
Next moveDraft the capture flow for one family recipe.
Same origin, same trail, new session

Continuing instead of starting over.

The next AI is oriented before the first turn.

Next session · with PCP
AIPicking up where you left off: memory-first capture, product framing parked. Next step is drafting the capture flow. Ready?
YouYes. Start with my grandmother’s stew.
Mid-thought on turn one.
Research foundation

Coherence becomes visible.

Purpose research asks how human intent can remain coherent when AI systems are fluent but not yet grounded. VDF turns that question into a visible interface, where the proof is in the purpose.

Purpose Research

AI Safety Infrastructure

Embedded a model-agnostic posture control system for conversational coherence.

Read the research
Formal Modeling

Theoretical Foundation

Applied systems modeling and simulation to intent-driven interaction.

See the model
Field Research

90-Day Field Validation

Validated coherence infrastructure in a live organization through artifact-first alignment.

View findings
Visual Decision Framework

Structure precedes output.

VDF gives AI-mediated thinking a visible structure people can see, adjust, and inspect, so ideas, artifacts, research, and decisions stay mathematically governed as they evolve.

Join the waitlist