Guides ·August 24, 2026

Air fryer food safety: why the internal temperature never changes, whatever the manual says

Every manual on this site gives you a time and a temperature dial setting: “180°C, 15 minutes.” None of that is a safety guarantee. It’s a cooking instruction, calibrated by one manufacturer, on one basket design, tested (or not) under conditions the manual never fully describes — see our guide to why cook times vary so much between brands for how far apart two official charts can land for the exact same food. The only number that actually tells you whether a piece of chicken, pork, or salmon is safe to eat is the internal temperature at the center, measured with a probe thermometer. On this site, that number comes from exactly one source, always: the USDA. Never a manufacturer manual, never an appliance-specific adjustment, never an average of the two.

The four thresholds, and why each one is what it is

The USDA’s Food Safety and Inspection Service (FSIS) publishes a small set of minimum safe internal temperatures that don’t change based on the cooking method, the appliance, or the recipe:

• Poultry — 74°C (165°F), no rest required. Whole chicken, chicken breast, thighs, wings, ground chicken or turkey — all the same figure. See FSIS’s chicken safety guidance for the source. It’s the number behind chicken breast and chicken wings on this site.
• Ground red meat — 71°C (160°F), no rest required. This is stricter than the whole-cut figure below because grinding can spread surface bacteria all the way through the meat, not just leave it on the outside. See FSIS’s ground beef guidance, which is the figure behind burgers on this site. The same logic sets 71°C for ground turkey and chicken too, one notch below the fresh-poultry rule only because it’s poultry, not because it’s ground — see FSIS’s sausage guidance.
• Whole cuts of beef, pork, veal, and lamb, plus fish — 63°C (145°F). Whole-muscle meat carries bacteria mainly on the surface, so a lower core temperature is safe provided the surface itself was properly seared or crisped. For beef, pork, veal, and lamb, FSIS adds a mandatory 3-minute rest after reaching 63°C, because the temperature keeps climbing slightly off the heat and the hold time itself helps finish off surface bacteria — see pork chops on this site. Fish doesn’t carry the same rest requirement; 63°C at the center, or flesh that’s opaque and flakes easily, is the target on its own — see salmon. Full chart on FSIS’s Safe Minimum Internal Temperature page.
• Pre-cooked, breaded, or reformed products — reheated to 74°C (165°F), regardless of the ingredient. Fish sticks and chicken nuggets both use this figure even though one is “fish” and the other is “poultry”, because neither is a raw protein being cooked for the first time — the fish or meat inside was already fully cooked at the factory before freezing. USDA and FDA guidance treats reheating a previously-cooked product as a different question from cooking something raw, and sets the same 165°F answer regardless of what the product is made from.

Bacon is the one common food on this site with no internal-temperature figure at all. It’s a cured and smoked product, and “done” for bacon is a texture judgement — crisp and golden — not a temperature target. So the bacon page shows no threshold at all, rather than an invented one.

Why the number never moves between appliances

Every food entry on this site has a safeInternalTempC field that’s set once, from USDA guidance, and never touched again when we add a new appliance or a new manufacturer measurement. It’s completely decoupled from the provenance field that tracks whether a cook time came from a manufacturer’s manual, our own testing, or a reader — see how we test, rule 4. A manufacturer is free to recommend whatever cook time and temperature it wants; the safety number we display next to it never changes because of that recommendation.

The one real drift we found: a Cuisinart recipe, 20°F under the line

We went looking for cases where a manual’s own safety language disagreed with USDA guidance, and found exactly one. A recipe inside the Cuisinart TOA-60 manual tells you to cook salmon to an internal temperature of 125°F (52°C) — twenty degrees Fahrenheit below the USDA’s 145°F (63°C) fish threshold. Read in context, this isn’t sloppiness: 125°F produces a medium-rare salmon, a texture plenty of cooks deliberately aim for with fresh, traceable fish, and it appears inside a specific recipe rather than as a general safety statement. The same manual gets chicken right four separate times elsewhere, consistently citing 165°F across several different recipes — so this isn’t a manual that’s careless about safety figures in general, it’s one recipe that assumes a level of judgement about fish quality that a printed number can’t convey. For comparison, the Ninja AF400UK quick-start guide’s own fajitas recipe calls for chicken cooked to 75°C — one degree above the USDA line, on the safe side of the threshold rather than under it.

That contrast is exactly why this site can’t take a safety number from any single manual, even a generally reliable one: the same document that’s correct four times can still drift once, in a context where a manufacturer is describing a preferred eating texture rather than a safety minimum. Whatever the recipe says, whatever texture a manual is aiming for, the number we display next to salmon on this site stays at 63°C/145°F — a cook is always free to stop earlier for their own taste, in full knowledge of where the safety line actually sits.

Most manuals don’t mention a safety number at all

Searching the full text of every manual in our corpus that has a usable cooking chart — 27 in total, across 12 brands — for any mention of an internal, safe, or core temperature turned up exactly seven that mention one in any form. The other twenty talk about safety only in terms of electrical or mechanical risk (hot surfaces, cord placement) and never once reference a cooking temperature target. That’s not a gap in our research; it’s the actual state of the manuals as written. It’s also the strongest argument for keeping safeInternalTempC as a field sourced exclusively from USDA guidance rather than from manuals: for the vast majority of appliances, there is no manufacturer figure to check in the first place, correct or otherwise.

The one tool that answers the question

A cheap probe thermometer — under £10 — is the only tool that actually answers “is this safe yet?” A timer answers a different question: “has this appliance finished its manufacturer’s programme?” The two questions usually agree, but not always, and the gaps documented in our guide to why cook times vary between brands show exactly how far apart two manufacturers’ timers can drift for the same food. When you use our calculator, the safety figure shown next to your result is the same number regardless of which appliance you picked or where its cook time came from — that consistency is deliberate, and it’s the one thing on this site that a brand’s own manual can never override.