NUTTAB vs USDA: Why Your AU Calorie App Is Probably Wrong

The Australian Food Composition Database (AFCD, previously NUTTAB) covers 1,588 foods built from Australian samples and reports up to 268 nutrients per food; USDA FoodData Central covers about 6,220 Foundation foods with up to 117 components from US supply. When a research team mapped Australian diets through both databases, individual fat intakes differed by more than 10 percent for 69 percent of participants. If your calorie app defaults to USDA, your AU food entries are wrong because the database is not built on the food on your plate.

Most calorie tracking apps used in Australia were not built in Australia. MyFitnessPal, Lose It!, and Cronometer all default to United States Department of Agriculture (USDA) data when their crowdsourced entries fall short, and the user-submitted entries themselves are weighted heavily toward US brands and serve sizes. The Australian Food Composition Database — maintained by Food Standards Australia New Zealand (FSANZ) and informally still called NUTTAB by many dietitians — is a different reference altogether. As of May 2026, AFCD is on Release 3.0 and contains analytical values measured on Australian wholemeal bread, Australian tinned tuna, Australian banana cultivars, and Vegemite — not their US equivalents.

This guide explains what AFCD is, where it diverges from USDA, where that gap actually matters for everyday tracking, and what an Australian user can do to keep their daily kilojoule numbers honest without ditching their app of choice.

What is the Australian Food Composition Database (AFCD), and how does it relate to NUTTAB?

AFCD is FSANZ's reference database of analytical nutrient values for foods sold in Australia. It is the same database that was published as NUTTAB from 1978 until 2019, when FSANZ renamed it to make its contents clearer. Release 3.0 covers 1,588 foods with up to 268 nutrients each and is the data source most Australian Accredited Practising Dietitians and food manufacturers rely on.

NUTTAB stood for "NUTrient TABles." Older editions — NUTTAB 1995, NUTTAB 2006, NUTTAB 2010 — were physically printed tables and downloadable spreadsheets. After the 2019 rename, FSANZ shifted the resource to a searchable web interface at afcd.foodstandards.gov.au. The underlying laboratory data did not change with the name; the rebrand was a labelling exercise. Many AU dietitians and university nutrition courses still say "NUTTAB" out loud and mean AFCD.

AFCD sits alongside two other FSANZ-maintained resources you will see referenced in food labels and apps:

If you want to see the difference yourself, the AFCD search page is free to use without registration. The same Granny Smith apple, the same Greek yoghurt, and the same Australian wholemeal loaf can return materially different energy and nutrient values than the corresponding USDA Foundation Food entries.

How is the USDA FoodData Central different from the AFCD?

USDA FoodData Central is the US Department of Agriculture's comprehensive food database, hosted at fdc.nal.usda.gov. It contains five data types — Foundation Foods, SR Legacy, Survey (FNDDS), Experimental, and Branded — and Foundation Foods alone cover roughly 6,220 commodity items with up to 117 components per food, sampled from US supply chains.

The two databases were built for different jobs. USDA has to serve a 330-million-person market with hundreds of thousands of branded products, so it leans on a larger Branded layer (USDA Branded Foods, updated monthly from GS1) and a relatively narrow set of high-quality Foundation items. AFCD has to serve a 26-million-person market with a tighter brand landscape, so it puts most of its budget into deep nutrient profiling — up to 268 nutrients per food versus USDA's 117.

For brand-specific items, the best calorie tracking apps comparison covers which apps have meaningful AU brand coverage. The general pattern: AU branded foods get accurate data when the manufacturer has either submitted to the Australian Branded Food Database or curated their own brand entry. Generic foods — "wholemeal bread," "cottage cheese," "banana medium" — are where the database choice matters most.

Does the AFCD vs USDA difference actually matter for everyday calorie tracking?

It matters more for individual day-to-day accuracy than for population averages. A 2013 Journal of Food Composition and Analysis study mapped the same Australian dietary intake data through both an Australian database and an extended US database and found macronutrient group averages differed by less than 5 percent for energy, carbohydrates, protein, and fibre — but at the individual level, fat intake differed by more than 10 percent for 69 percent of participants, and folate by 96 percent.

That study, by Roy and colleagues, fed the same food records into two database backends. At the group level, the two databases looked interchangeable for energy (3 percent difference), protein (5 percent), carbohydrates (-2 percent), and fibre (3 percent). Fat was already different by 11 percent at the group level. At the individual level the discrepancies were much larger:

A few patterns make this worse for Australian users specifically:

• Crowdsourced entry quality. A 2024 sampling review of nutrition app databases found MyFitnessPal entries had a 23.1 percent error rate against verified reference values. Foods marked with the green verified tick fare better, but the default search result is often a user-submitted entry.
• Brand-name collisions. "Tim Tam," "Vegemite," "Anzac biscuit," "Milo," and "Weet-Bix" all return multiple conflicting entries, some sourced from US import-store data using imperial serve sizes that do not match the AU pack.
• Cooked-versus-raw drift. AFCD records many ready-to-eat AU staples — grilled lamb chop, baked snapper, white rice cooked — using AU recipe conventions. USDA records the same nominal foods using different oil, salt, and water ratios, which can shift energy by 5-15 percent per serve.
• Energy unit translation. AU labels use kilojoules as the primary unit (1 kcal = 4.184 kJ). Apps that store data internally in kcal and convert on the fly can round at the conversion step, accumulating 20-40 kJ of drift per logged item.

How do you get more accurate calorie tracking on AU foods?

You get more accurate tracking by anchoring AU-specific foods to AFCD or the AU label whenever possible, by scanning AU barcodes instead of typing food names, and by treating any generic food entry without a verified flag as a ±15 percent estimate. The most precise option for Australian foods is to enter values manually from the on-pack Nutrition Information Panel, which by law is calculated through FSANZ's Nutrition Panel Calculator.

A practical AU-tracking checklist:

For ethnic cuisine and dishes whose ingredients diverge from both AU and US databases, see the global cuisines calorie tracking guide. For mixed plate accuracy specifically, the AI vs manual tracking comparison covers when photo logging beats database lookup.

Why don't more calorie apps use the AFCD by default?

Most calorie apps default to USDA data because USDA FoodData Central is free, has the largest single set of measured Foundation Foods, and ships with a stable API that any developer can plug into in an afternoon. The AFCD is also free, but its smaller food set, lower update cadence (a major release every 3-4 years), and lack of a programmatic API have historically made it harder to integrate at scale.

The commercial calorie-app market is overwhelmingly US-built and US-funded. MyFitnessPal was founded in San Francisco. Lose It! is from Boston. Cronometer is Canadian but draws heavily on NCCDB (a US dietetic reference). The product teams that built these apps had USDA on hand and Australian data sitting one ocean away in a different format.

The result is a structural blind spot. Australian users get a perfectly usable app for tracking calories in general, but the underlying numbers come from a food supply they are not eating. KCALM, built in Sydney, prioritises AFCD-aligned values for Australian foods identified by AI vision and falls back to USDA only when the AU reference is absent. This is the same gap the psychology of calorie counting guide flags as the root of "my app says I am in a deficit but I am not losing weight" frustration — the database can be quietly wrong before any user error enters the picture.

The fix at the user level is the seven-step checklist above. The fix at the industry level is for AU regulators and FSANZ to publish a modern AFCD API and for app developers to ingest it. Neither has happened yet as of May 2026.

Frequently Asked Questions

Is NUTTAB still updated, or has it been replaced?

NUTTAB was renamed, not replaced. FSANZ rebranded NUTTAB to the Australian Food Composition Database (AFCD) in 2019 to make the contents clearer to non-technical users. The underlying lab data is the same continuous research programme. Release 1 of AFCD was published in 2019 with 1,534 foods, Release 2 in 2022, and Release 3.0 is current as of May 2026 with 1,588 foods and up to 268 nutrients per food.

Does MyFitnessPal use the AFCD for Australian users?

Not by default. MyFitnessPal's primary database is crowdsourced user submissions backed by USDA reference data, with no automatic switch based on user country. AU users can find AFCD-aligned entries through verified (green tick) submissions and through Australian Branded Food Database barcode scans, but generic food searches typically return USDA-style entries first. A 2024 sampling review found roughly 23 percent of MyFitnessPal entries had measurable errors against verified reference values.

How accurate is the kJ figure on Australian food labels?

The kJ figure on an Australian Nutrition Information Panel is calculated via the FSANZ Nutrition Panel Calculator from a recipe ingredient list, then rounded to comply with the Food Standards Code. It is not laboratory-measured per pack. The label is allowed to vary from the actual product by a small percentage and is best treated as accurate to within ±10-15 percent of the food in the packet. For chain restaurant kJ labelling specifically, see the AU chain calorie guide.

Why does my Vegemite show different calories on US vs Australian apps?

Because the US-side entry is usually scraped from an import-store website using a different serve size and sometimes an older formulation, while the Australian-side entry traces to the AU Vegemite label, which is calculated via NPC from AFCD ingredient values. The actual product is the same in either market, but the entry chain in the database is different. Scan the AU barcode on the jar to force the AU entry.

Can I just enter foods from the AFCD website manually?

Yes, and for whole foods this is the most accurate option short of laboratory analysis. The AFCD search at afcd.foodstandards.gov.au is free, requires no account, and returns per-100 g values in both kJ and kcal alongside the full nutrient breakdown. Most calorie apps let you create a custom food from per-100 g values in under a minute. This is the same workflow registered dietitians use when accuracy matters more than logging speed.

How does KCALM handle AU foods differently from US-built apps?

KCALM's AI vision pipeline is trained to recognise AU-common foods like Vegemite on toast, Anzac biscuits, lamingtons, Tim Tams, flat whites with full-fat milk, and Australian-portion meat servings, and it prioritises AFCD-aligned reference values for those identifications when available. The app falls back to USDA only when there is no AU reference for an identified food. For new users, the getting started guide walks through how to verify your first few logged AU foods before trusting the numbers in calibration mode.

Sources