#coffee

Coffee-roasting maths as an API, computed locally and deterministically — the roast-profile numbers a home or specialty roaster tracks batch to batch. The loss endpoint works the weight-loss relationship from any two of the green weight, the roasted weight and the loss percentage: weight loss % = (green − roasted) / green × 100, so 1 kg of green dropping to 840 g is a 16 % loss, a 15 % target leaves 850 g, and to bag 800 g of roasted you charge 952 g of green (roasted ÷ (1 − loss%)). Light roasts shed about 12–14 %, medium 15–17 %, dark 18–20 %. The development endpoint computes the development time and the Development Time Ratio (DTR) from the total roast time and the first-crack time — DTR = (total − first crack) / total × 100, with most roasters aiming for roughly 20–25 %; times accepted as seconds or mm:ss. Everything is computed locally and deterministically, so it is instant and private. Ideal for coffee-roasting, roastery, specialty-coffee and roast-logging app developers, profile and batch tools, and roasting education. Pure local computation — no key, no third-party service, instant. Weights in grams, times in seconds or mm:ss. Live, nothing stored. 2 compute endpoints. This is roast-profile maths; for brewing ratios use a coffee-brewing API.

api.oanor.com/coffeeroast-api

Coffee brewing maths as an API, computed locally and deterministically. The ratio endpoint works out a brew recipe from any two of the coffee dose, the water and the brew ratio — water = coffee × ratio — and reports the third value, the ratio as 1:N, the number of cups and whether the recipe sits around the SCA "golden ratio" of about 1:16–1:17. The espresso endpoint does the same for espresso from any two of the dose, the yield and the brew ratio (yield = dose × brew ratio), labelling the shot ristretto, normale or lungo. The extraction endpoint computes the extraction yield, EY% = (beverage mass × TDS%) ÷ dose, from the dose, the brewed beverage mass (or the water, estimating the mass the grounds retain) and the measured total dissolved solids, then classifies the brew as under-extracted, ideal or over-extracted and weak through very strong against the SCA brewing control chart. Masses are in grams, water in grams or millilitres. Everything is computed locally and deterministically, so it is instant and private. Ideal for specialty-coffee, café, brewing-scale and recipe app developers, pour-over and espresso tools, and barista training. Pure local computation — no key, no third-party service, instant. Live, nothing stored. 3 endpoints. This is coffee brewing maths; for cooking-unit conversions use a cooking API and for caffeine intake use a caffeine API.

api.oanor.com/coffee-api

Caffeine metabolism maths as an API, computed locally and deterministically with a first-order (exponential) decay model. The level endpoint computes how much caffeine remains in the body after a given time from a dose and a half-life (about 5 hours by default), as milligrams and a percentage, and how long until it falls to a chosen threshold. The timeline endpoint returns an hour-by-hour decay curve and the time until caffeine is "sleep safe" — below a threshold (50 mg by default) — handy for working out a coffee cut-off before bed. The sources endpoint gives the typical caffeine content of common drinks (brewed coffee, espresso, tea, energy drinks, cola and more) for a single drink, or totals a list such as two coffees and a cola. Everything is computed locally and deterministically, so it is instant and private. This is informational only: real caffeine half-life varies widely between people (roughly 3–7 hours, and much longer in pregnancy or with certain medications) — it is not medical advice. Ideal for coffee, sleep and wellbeing apps, energy-drink and habit trackers, and quantified-self tools. Pure local computation — no key, no third-party service, instant. Live, nothing stored. 3 endpoints. This is caffeine pharmacokinetics; for a drug reference database use a drug API.

api.oanor.com/caffeine-api