BayesianBrewBETA
Sign in

Dial in faster. Every new bag.

Bayesian Brew is a coffee brewing optimiser. Rate a few brews and a Bayesian model narrows whichever parameters you care about — grind, dose, ratio, temperature, bloom, anything you can dial — converging on your taste, without wasting half a bag.

01

Create a brew profile

Tell us about your bag, gear, and which parameters you want to dial.

02

Brew the suggested recipe

We pick the next one to try. You brew it.

03

Score it, and repeat

Rate the brew. We learn, suggest again, and converge on your taste.

The loopbrew 1 / 10
score4/ 10your rating
1471012345678910
Try nextexploring · expect surprises
Grind=2.20
Temp=92°C
Ratio=1:15
Why thisparameter importance
Grind0.30
Temp0.28
Ratio0.22
Strongest interaction
Grind × Ratio

Still figuring out which dial matters most.

Free during beta. No card. We'll email a sign-in link.

Tuned to your palate.

Every rating sharpens what we suggest next. The dial-in is yours, not borrowed from someone else's beans.

Meets you mid-recipe.

Start from whatever you'd brew today, and vary the parameters you actually want to play with. It moves from there.

Shows its working.

See which parameters are pulling the score, and how they interact. No mystery suggestions.

Tells you when to stop fiddling.

Once the recipe stops moving, we say so. Drink the rest of the bag in peace.

One dial at a time.

Slice the model along a single parameter to see how the predicted score moves — handy when you want to nudge grind without touching anything else.

How two dials play together.

Two-parameter heatmaps show where your best brews cluster — the kind of interaction that's hard to spot by feel.