The dataset

Everything this site publishes is available as plain JSON. 38 air fryers across 15 brands, 18 foods, and 289 cook time measurements — each one carrying the manual or product page it was read from, and the date it was retrieved.

The data is released under CC BY 4.0: use it, quote it, build on it, as long as Right Cook Time is credited. No key, no rate limit, no sign-up. The endpoints are rebuilt on every deployment, so they never drift from the pages.

Endpoints

• /api/appliances.json · 38 records
  One record per appliance: brand, model, model code, regional variants, wattage, capacity, basket type, maximum temperature, preheat, and the sources the specs were verified against.
• /api/foods.json · 18 records
  One record per food: reference temperature and time, reference weight, the USDA safe internal temperature, prep and cooking steps, doneness test, and sources.
• /api/measurements.json · 289 records
  One record per appliance × food pair: temperature, time range, weight range, provenance, manual page number where available, measurement date, and sources.
• /api/dataset.json
  The schema.org Dataset descriptor itself, listing the three distributions above with their live counts.

What provenance means

Every measurement is labelled, and the label is shown on the page as well as in the JSON — never hidden in the source. Right now the corpus breaks down as:

• manufacturer — 289 measurements

manufacturer means the value comes from the brand's official chart and has not been independently re-measured here. tested means measured on the appliance.reader means submitted by a reader and published after moderation. See how we test for the full method.

Worth saying plainly, since the list above says it anyway: no reader row exists yet. Everything here can be reproduced by anyone willing to read the same manuals, and that is a fair thing to hold against the dataset. The rows that couldn't be reproduced — someone cooking a specific food in a specific machine and timing it — start arriving when readers send them in.

Brands covered

Bosch, Breville, Chefman, Cosori, Cuisinart, Daewoo, Gourmia, Instant, Midea, Ninja, Philips, Russell Hobbs, Tefal, Tower, Typhur. Browse them from the appliance index, or compare two brands directly from the comparison hub.

Limits worth knowing before you cite it

• Most values are manufacturer figures, not independent measurements. Two manuals genuinely disagree on the same food more often than you would expect — that disagreement is preserved, never averaged away.
• Cook times are never averaged across models of the same brand: a 1000 W and a 2600 W appliance keep separate records even when the brand markets them as the same product line.
• Safe internal temperatures come only from the USDA, never from a manual. Where a manual disagrees with the USDA, the USDA figure is the one shown.
• Coverage is UK-first and incomplete: not every appliance has a measurement for every food.