Labels are data
A label is a set of regulated statements, not a picture. Treating it as structured data is what makes completeness checkable and translation reusable.
About Correling
Label content deserves the same rigour as the recipe
Correling grew out of years of running multilingual food-label translation work - and watching the same content be retyped, retranslated and re-checked for every market.
Food producers who export face a quiet, expensive problem: the content of a product label - ingredients, allergens, nutrition, storage, producer details - has to be exactly right in every market language, and it changes more often than anyone plans for.
Correling was built to run that workflow properly. The platform has managed hundreds of product labels and over a thousand field translations across dozens of label templates, growing out of a production system used by real exporting producers.
The core idea has stayed the same throughout: model the label as structured fields, translate each field once, compose labels from templates, and publish a single authoritative page per label.
Principles
What we optimise for
Translate once, reuse everywhere
Terminology drift is a quality bug. Approved translations belong to fields, not to individual documents, so consistency is automatic rather than heroic.
Traceability over trust
When an inspector or retailer asks why a label says what it says, the answer should be a history log - not an archaeology project through inboxes.
Keep what works
Teams have years of label content in old systems. Migration tooling that preserves that history is part of the product, not an afterthought.