Tooling
Live

Food Databases & Open Food Facts

Open Food Facts, USDA FoodData Central, CIQUAL — the open and reference datasets behind every nutrition app.

The data layer beneath everything else in this section. For consumer-facing nutrition trackers see Nutrition Tracking — Cronometer & MyFitnessPal; for recipe-side imports see Recipe Import & Scrapers; for pantry / barcode scanning workflows see Pantry Inventory & Grocy; for allergen overlays see Allergen & Intolerance Tracking.

Open Food Facts (the OSS dominant)

  • ★ ★ Open Food Facts — Wikipedia-of-food; ~3M product entries with barcode + ingredients + allergens + Nutri-Score + NOVA + image. Crowdsourced, auditable, CC0. Mobile apps for iOS / Android. Mature API; daily database dumps. The default nutrition data layer for any new app. AGPLv3 server.

What it covers:

  • Branded grocery products by EAN / UPC barcode.
  • Nutri-Score (A–E) — French-origin front-of-pack rating.
  • NOVA classification — ultra-processed-food rating (1 = unprocessed, 4 = ultra-processed).
  • Allergen tags — gluten, milk, peanuts, etc., with cross-contamination flags.
  • Eco-Score — environmental rating; methodology is debated.
  • Ingredients OCR from product photos via the Robotoff ML pipeline.

The community has been on a tear since 2023; data quality is now competitive with paid databases for most categories. Yuka, several big retailers, and Cronometer's branded section all depend on it.

USDA FoodData Central (the gold-standard reference)

  • ★ ★ USDA FoodData Central — free, unlimited API after key registration; the canonical reference for raw / minimally-processed food nutrition. Five sub-datasets:
    • SR Legacy — the classic Standard Reference; ~7,800 foods; deep nutrient profiles.
    • Foundation Foods — newer; modern lab analyses; ~1,000 foods, growing.
    • FNDDS — Food and Nutrient Database for Dietary Studies; mapped to NHANES survey codes.
    • USDA Branded — ~400k branded products via GS1 Global Data Synchronization; DV-percentages baked in.
    • Experimental — research datasets.
    • https://fdc.nal.usda.gov

USDA FDC is the right call for "what's in 100g of plain chicken thigh." Open Food Facts is the right call for "what's in this Trader Joe's bar I just scanned."

Regional reference datasets

  • CIQUAL (France) — ANSES; rigorous; French-language UI; download as CSV; widely used in EU nutrition research.
  • CNF (Canada) — Canadian Nutrient File; comparable to SR Legacy.
  • McCance & Widdowson (UK) — Public Health England; UK reference; paid in some forms.
  • NEVO (Netherlands) — Dutch food composition; paid for redistribution.
  • AUSNUT — Australian.
  • NIN-IFCT (India) — Indian Food Composition Tables.
  • FAO INFOODS — global meta-database aggregating regional sources.

Specialty / scientific databases

  • Phenol-Explorer — polyphenols by food; academic; free.
  • DTU FOOD (Denmark) — heavy metals + contaminants.
  • EuroFIR — European federation of food databases; paid commercial access.
  • NCCDB — Nutrition Coordinating Center; paid academic; powers Cronometer's research-grade rows.
  • ESHA — paid commercial; powers many enterprise nutrition apps.
  • Glycemic Index Foundation DB — GI / GL values by food.

Allergen / ingredient databases

  • ★ Open Food Facts allergens — covers EU 14 + US Top-9; per-product flags + cross-contamination warnings.
  • FDA / FSIS recall feeds — free RSS / JSON; tap if you're building a food safety app.
  • ALLERGEN ONLINE (FARRP / Univ Nebraska) — paid academic; allergen protein database.

API / dataset access patterns

  • Open Food Facts:
    • REST API: world.openfoodfacts.org/api/v2/product/{barcode}.json — no key needed.
    • Daily MongoDB / Parquet / JSONL dumps for bulk analysis.
    • Robotoff API for ML-derived classifications (categories, brands, nutritional inference).
  • USDA FDC:
    • REST API; free key; modest rate limit (1,000 req/hr).
    • Bulk CSV / JSON downloads of each sub-dataset.
  • CIQUAL — CSV download; no API.
  • Most others — CSV downloads; license-restricted for commercial use.

Building on these

  • Mealie / Tandoor use OFF for product lookups in shopping lists.
  • Grocy + barcode scanner + OFF userscript — see Pantry Inventory & Grocy.
  • Yuka — OFF data + opinionated rating overlay.
  • Cronometer — combines USDA FDC + NCCDB + OFF for branded.
  • Edamam, Spoonacular — proprietary databases that lean on USDA + OFF data internally.

License watch-outs

  • Open Food Facts: data CC0 (public domain); server AGPLv3; logos and product photos are user-submitted (assume CC-BY-SA, attribute).
  • USDA FDC: US Government work, public domain in the US.
  • CIQUAL: Open license / Etalab; redistribution-friendly.
  • NEVO, McCance & Widdowson, ESHA, NCCDB: commercial / restricted; don't redistribute.
  • The Branded subset of USDA FDC comes from GS1 GDSN; verify usage rights for commercial scenarios.

Practical guidance

  • ★ Open Food Facts first; USDA FDC second; commercial DB only when forced. Most consumer apps never need anything paid.
  • Cache aggressively. OFF is donation-funded — don't hammer the API. Use the daily dumps for backfill.
  • Contribute back. If you scan a missing product, OFF's mobile app makes adding it trivial.
  • Allergen flags are not legally compliant on their own. Always label "may contain" with a regulatory disclaimer if you ship a real product.
  • Branded data drifts. Reformulations happen; don't trust a 2018 row from USDA Branded for today's macros.

Pick this if…

  • Default barcode + branded: Open Food Facts.
  • Default raw-ingredient nutrition: USDA FoodData Central.
  • EU regulatory work: CIQUAL + Open Food Facts allergens.
  • Research-grade micronutrients: USDA Foundation Foods + NCCDB (paid).
  • Polyphenol / phytochemical detail: Phenol-Explorer.
  • You need a hosted "give me everything": Spoonacular or Edamam — see Recipe Import & Scrapers.

Previous

Culinary Education & Courses

Next

Ingredient Substitutions

On this page