Methodology
How we evaluate air fryers
This is our headline promise: a method you can check. Everyone in this category says they tested twenty air fryers. We have not lab-tested any of them, and we say so — here is what we do instead.
0
Air fryers we claim to have lab-tested
0
Sponsored placements accepted
0
Free products accepted for coverage
Those zeros are the point, not a gap. We would rather publish that we haven't tested something than pretend we have. What we bring instead is a consistent, reproducible method applied the same way to every machine.
1. How we source the picks
We start from what people actually search for and what Amazon actually stocks — because a recommendation you cannot buy is useless. We build a longlist from the category (the well-known models, the cult favorites, the value options), confirm each one is currently sold and matches the brand and model we intend, and then cut it down. Nothing gets a spot because a brand asked; brands cannot buy a spot, because we accept no sponsored placements.
2. The scoring rubric
Every machine is scored out of 10 against a small set of criteria chosen for the category — usually capacity, crisping, ease of use, materials, running cost, and value (roundups pick the four to six that matter most for that page). The overall score is the mean of those metric scores, shown to one decimal place. We publish the sub-scores on every card so you can see exactly what the number is made of.
The score is a judgment from documented research — published specs, manufacturer documentation, materials, and aggregated owner reviews. It is not a measurement we took in a lab, because we do not have one. This is also why you will never see an aggregate star ratingfrom us in a page's structured data: we have no customer reviews of our own to aggregate, and inventing one would be a lie.
3. The running-cost calculation (our signature)
The number no competitor publishes, and the reason this site exists. For every machine we compute what a typical cook costs using a formula anyone can reproduce: cost = (watts ÷ 1000) × (minutes ÷ 60) × price per kWh. We take the wattage from the manufacturer's stated rating, use the U.S. Energy Information Administration's average residential electricity price as the default rate, and show our assumptions on the page. You can run your own numbers — your machine's wattage, your utility's rate — on our running-cost calculator. Where a manufacturer doesn't print an exact wattage, we say we're using a typical figure for the class rather than inventing a precise one.
4. How prices work
Every price on this site is pulled live from the Amazon Product API and stamped with the date it was fetched. If our data is more than 48 hours old, the number disappears and the button falls back to "Check price on Amazon" — we would rather show you no number than a stale one. We never type a price into the page by hand. Amazon's price at checkout is always the one that counts.
5. Materials and safety claims
Whether a basket is PTFE nonstick, PFAS-free ceramic, stainless, or glass is a verifiable material fact, and we treat it as one — read off the manufacturer's own listing, never guessed. Any health or safety statement is traced to a primary source (a food-safety agency or the manufacturer), which is why our are air fryers toxicguide cites the FDA and EPA rather than a competitor's article.
6. The "don't buy this if" rule
Every machine we recommend carries a line telling you who it is notfor. It is the fastest way we know to prove we are on your side rather than the retailer's, and no competitor page we studied does it. Every roundup also names at least one pick to skip.
7. Updates and corrections
Prices refresh continuously; roundups are reviewed at least quarterly and safety guides re-checked against source updates. Every page shows a visible last-updated date. If we get something wrong and you tell us, we correct material errors within 48 hours and note the correction on the page. See our editorial policy for the full standards, and contact us to flag anything.