Roadmap / Proof before scale
Olfaction is the wedge because it is the shortest path from memory to feeling. Scent is where Scentient starts. It is not where Scentient ends. Each horizon below is what we build on top of the loop closed in the horizon before.
Thirty years of named scent choice gives Scentient a physical vocabulary and a rare cold-start advantage.
An early capture direction: authored memories, spatial atmosphere edits, and candidate comparison as structured signal.
Custom fragrance is the first artifact: tangible, shippable, and precise enough to learn from.
Accepted refinements and later responses can improve future memory-to-scent translations.
The first consumer surface for memory-to-scent translation ships. The Demeter Fragrance Library rail handles physical fulfillment. The loop starts with authored memory, candidate comparison, and accepted refinements.
The Scent DNA profile and the translation surface are exposed to selected developers and design partners. First B2B integrations land in hospitality, retail, and immersive entertainment. The graph crosses its first scale milestone.
Atmosphere-aware environments begin using Scentient as a sensory direction layer. Hardware partners can route emission decisions through the graph when scent should become ambient. The category becomes legible beyond fragrance.
Olfaction was the wedge. The graph can extend to ambient sound, directional light, haptic vocabulary, and thermal cues. Scentient becomes a semantic layer for sensory interfaces that need memory-conditioned direction.
The long-term aim is for environments, agents, and interfaces to use the graph as a sensory-affective layer when deciding what to emit, when, and why. The infrastructure ambition remains the same; the claim stays earned.
The Ask / For design partners
World models and embodied AI start from the senses. The world's olfactory data is “smells good versus smells bad” — the upside-down logic Demeter Fragrance Library has been quietly fixing for thirty years. Transformer-class AI just made the four-hundred-receptor combinatorial problem tractable. The window between solvable and solved is small.
The only team with both the labeled olfactory archive — thirty years inside Demeter Fragrance Library, organized around scent-memory — and the computation layer that turns it into a graph. Demeter Fragrance Library is the trust layer, the fulfillment rail, and the head start no capital can buy in a decade.
A machine-readable substrate for scent-memory: memory-to-scent translation, an affective memory graph, an affective value function, and closed-loop affective learning. The opportunity starts with scent as the first physical proof surface and expands as more interfaces need memory-conditioned sensory direction.