Method & Docs
How Cosma works
This page documents how scans are processed, how ingredients are normalized, and how context is presented. The goal is clarity and provenance, not scores or judgments.
Core commitments
- Transparency over persuasion
- User control over data
- Minimal, purpose-driven collection
- No dark patterns or hidden monetization
1) Scan or input
You scan a barcode or enter an ingredient list. The raw input is stored as part of your personal record so you can trace back to the source.
Image placeholder
Image placeholder
2) Normalize ingredients
Ingredients are normalized against Cosma’s internal knowledge base. Synonyms and formatting noise are reconciled so the record stays consistent over time.
3) Context, not verdicts
Each ingredient is presented with usage context, classifications, and notes. Cosma does not apply scores or “good/bad” labels.
Image placeholder
What we capture
- Ingredient lists and barcode inputs you provide
- Normalized ingredient names and common synonyms
- Context tags (usage, classification, concerns)
What we don’t do
- No medical diagnosis or treatment advice
- No cross-app tracking or ad targeting
- No scoring systems or verdict labels
Privacy and data handling
Accounts are optional and only required for syncing saved items across devices. Data collection is minimal and purpose-driven.