Questions you might have

If yours isn't here, drop us a line at hello@xbrain.dev.

What does Palate actually do?

Predicts how much you'll personally like a product — cannabis, wine, beer, coffee, or a dish — before you buy or order it. It's one app that covers all five categories; the more you rate, the better the predictions get.

How is this different from Leafly / Wine Spectator / generic review sites?

Those tell you what the average person thinks. X-Brain learns *you*. Two users can look at the same product and see different predicted ratings — that's the point.

Do I need to be at a partnered dispensary or shop?

No. You can rate anything you've tried, anywhere. Live menu integrations are a convenience layer for ranked discovery — not a requirement for the model.

What signals does the model use?

For cannabis: 40+ terpenes, cannabinoids (THC/CBD/CBG/CBN), product format (flower/vape/edible/concentrate), brand, strain genealogy, and your effect/mood tags. For wine: varietal, region, vintage, ABV, pH, tannins, body, and similar effect-style tags.

How does the photo identification work?

Vision waterfall: COA QR scan first (fastest, free), then terpene-sticker OCR, then label OCR + brand match, then a vision LLM call against the local catalog. Falls back gracefully — you can always pick from a list.

Is my rating data private?

Your individual ratings are private to you and the friends you choose to share with. We aggregate anonymized data for partner dispensaries (e.g. 'users in your area predicted ≥4.5 on these items'). Your email and identity never leave the platform.

Do I need a different account for cannabis, wine, beer, coffee, and food?

No — one signup covers all five. Friends, badges, and the cross-brand Knowledge Graph span every category. The personal taste model is per-category (terpene preferences don't transfer to tannins or hop bitterness), but your account, social graph, and rating history are shared.

Is it one app or one per category?

One app. myPalate covers cannabis, wine, beer, coffee, and food — switch the active category anytime from your profile, and the whole experience (what to scan, what to filter, what you see) re-themes to it. One download, one account, one shared backend.

What's a cohort?

A cluster of users with similar taste fingerprints. New users borrow predictions from their cohort until they've rated ~20 products on their own — that's how the cold-start works.

Where does the catalog data come from?

Per category: cannabis from storefront and dispensary APIs (Dutchie, iHeartJane, Weedmaps, SweedPOS, Trulieve, Apothecarium); wine and food from Open Food Facts and USDA; coffee and beer from roaster and brewery catalogs. Live menus refresh nightly. We don't scrape sites that ask us not to.

Palate — personalized taste predictions