23 interactive arguments.
skincare marketing sells one number. goodbadgray measures the grays — dose, formulation, skin type, life stage. each demo below is one argument of the pitch, built to open from disk, offline, on real engine output. click any card.
“a little bit of white, a little bit of black, and lots of those damn grays.”
real data harvested read-only from the live gbg engine · 2026-07-08 · simulated elements labeled “illustrative” in-ui
short on time? six stops.
the insight → the engine → the person → the wedge → the platform → the money. about two minutes per demo; each one has a why this demo drawer with its investor thesis.
full arc: five acts, 23 demos, below. everything opens by double-clicking index.html — no wifi, no install, no build.
the science
the engine, proven: context-dependence as the product, a 14-axis scoring instrument, the dossier moat, and concentration estimation with honest error bars.
shades of gray
the brand manifesto as an interactive: one ingredient, one dial, and the verdict slides between good and bad as context changes.
drag one slider and the “isn’t this just another scoring app?” objection dies — context-dependence is the product, not a feature.
open demo →anatomy of a score
the 14-metric vector, dissected: what the engine measures, why 14 axes beat one number, and which real ingredients define each extreme.
turns the scoring model from a diligence liability into the company’s most defensible artifact.
open demo →the dossier
an encyclopedia page a chemist would respect: 89 real ingredient dossiers with mechanisms, confidence levels, life-stage flags, and environmental ledgers.
the data moat, browsable — bloomberg-for-ingredients, with 8,595 more dossiers in the live engine.
open demo →product x-ray
paste-a-label theater: a real INCI list decodes itself into roles, concentrations with confidence intervals, and a final verdict.
the engine estimates what brands won’t disclose — per-ingredient concentrations with monte-carlo confidence intervals.
open demo →the concentration engine
how do you know what’s inside? the Monte-Carlo concentration estimator, visualized ingredient by ingredient.
the hardest technical claim in the company, shown as disciplined statistics instead of a black box — for the technical partner in the room.
open demo →the person
same chemistry, your skin: transparent personalization math, life-stage verdicts, routine refereeing, and the shareable shelf moment.
one product, six faces
the same real sunscreen scored for six real profiles — six different verdicts, and the weight math shown in full daylight.
personalization as auditable math over real config vectors — the property that makes the engine sellable b2b.
open demo →nine months
a shelf of real products walks through conception to breastfeeding, and the verdicts flip stage by stage.
the most underserved, highest-willingness-to-pay segment in skincare — already served by the live engine, trimester by trimester.
open demo →routine architect
drag real products into an AM/PM routine and watch the engine referee the chemistry between them.
from checker app to operating system for the shelf — routine intelligence is the retention loop.
open demo →the interaction web
the engine’s knowledge of ingredient chemistry, drawn as a living force-directed constellation.
not a list of scores — a knowledge graph that compounds superlinearly with every new analysis.
open demo →shelf x-ray
a bathroom shelf of real products, x-rayed in one sweep — verdict chips, flags, and the one product to retire.
the shareable aha: your own shelf, re-scored around your own skin — and the hinge into commerce.
open demo →the market
where the money and the pressure live: dupes, claim audits, market maps, brand ratings, environmental regulation, and clinical white space.
dupe radar
same chemistry, a third of the price: vector similarity finds real dupes among real products.
the most viral behavior in beauty, generated mathematically instead of by influencer anecdote — with a commerce take on every swap.
open demo →claim vs reality
what the label promises vs what the ingredients measure — marketing language audited by the engine.
the accountability layer that sometimes sides with the brand — which is exactly why it earns trust and headlines.
open demo →market lens
every product plotted as price versus measured quality — and the brand-premium myth collapses on contact.
the b2b data-licensing slide disguised as a consumer chart: measurable quality prices the whole market.
open demo →brand report card
brands graded like public companies: portfolio-level scoring, flag rates, and a greenwash gap index.
how a data startup becomes a ratings authority the industry has to answer to — recurring b2b revenue included.
open demo →the reef ledger
what your sunscreen does after it leaves your skin: reef toxicity, persistence, and the regulatory wave, on real data.
regulation is the forcing function that turns transparency into compliance spend — and the engine already flags all of it.
open demo →medication guard
the first skincare layer that knows what’s in your medicine cabinet: real drug-interaction flags flip a real shelf.
clinical white space no competitor’s data model can even represent — the wedge into telehealth and pharmacy.
open demo →the platform
the compounding asset: network effects, the api, the magic moment, the pipeline, the telemetry, and the roadmap you can touch.
skin twins
the network-effect flywheel: profiles cluster into skin twins, and every scan makes the next recommendation smarter.
the network effect made tangible instead of asserted — every new user improves the priors for their twins.
open demo →api atelier
goodbadgray as infrastructure: the developer portal for the skincare intelligence API.
plaid for skincare: the product is already api-shaped, with real dossier json streaming out of a sellable primitive.
open demo →scan theater
the app’s magic moment, staged for a big screen: point at a label, get the truth in four seconds.
the four-second magic moment, choreographed for demo day — no wifi, no live camera, all real verdicts.
open demo →pipeline observatory
mission control for the analysis engine: one real product flows through every stage, all intermediate numbers real.
a disciplined multi-stage pipeline with inspectable state — not one opaque llm call. built for enterprise slas.
open demo →growth telemetry
the data asset as a dashboard: what the engine holds today, how it compounds, and the market it prices.
the honest traction slide for a data company: the asset is the traction, and the moat widens with every dossier.
open demo →futures gallery
four playable postcards from the roadmap: formulation copilot, skin journey, regulatory radar, and the mirror.
vision you can handle: every future sku runs on the same 14-metric spine already in production.
open demo →