Skip to content
Moleculo

Sources

Sources and methodology

Every fact about an additive comes from an official source, with a link and a date. What we can't cite, we leave empty and flag.

Method

  • The data is prepared when the site is built, not in your browser. The site makes no third-party requests while you use it.

  • Names, EU list sections, mandatory labelling statements and food categories are extracted automatically from the official texts, in Romanian and English, without rewording.

  • Molecular structures come from PubChem and are accepted only when the molecular formula matches the formula in the EU specification. For mixtures and polymers we don't show an invented molecule.

  • Nothing is generated by AI: we have no texts about health effects and we give no scores.

Coverage

  • E-numbers on the EU list340
  • with an EU specification338 / 340
  • with a definition (origin)184 / 340
  • with authorised food categories324 / 340
  • linked to EFSA opinions220 / 340
  • with a verified molecular structure169 / 340
  • with a functional class (colours and sweeteners only)63 / 340
  • with a plain-language explanation written by us0 / 340

Sources

Retrieved:

Software and fonts

  • Tesseract.js 7 + Tesseract LSTM (Apache 2.0), traineddata ron/eng 4.0.0 best_int (Apache 2.0)
  • three.js, React Three Fiber, drei (MIT)
  • MiniSearch (MIT), zod (MIT), zustand (MIT), idb-keyval (Apache 2.0)
  • Inter, Manrope, JetBrains Mono (SIL Open Font License 1.1)