No conversation loaded

Load an example or paste a conversation to begin.

G_user0n · 0e

The user's problem formulation will appear here as the conversation unfolds.

G_model0n · 0e

The model's inferred formulation will appear here, including implicit assumptions.

Substrate

Make human–GenAI collaboration legible.

Substrate extracts a paired, time-indexed representation of human–LLM dialogue: the user's evolving problem formulation, the model's inferred formulation, and the moments where they silently diverge.

Dual graph

Two typed problem graphs evolve in lockstep — one per side of the dyad.

Turn-by-turn

Watch nodes appear, get refined, or supersede each other as the conversation unfolds.

Divergence channel

Grounding acts, scope drift, ignored constraints, premature commitment — all surfaced.

Start with a curated example

Extraction is intentionally rough; the contribution is the representation, not the F1 score. Once an externalised, inspectable dual representation exists, the moments that were previously invisible become legible.

How it works

Experimental research prototype, provided as-is with no warranty. By using Substrate you agree to the Terms of Use.