INCI FastScan: Guide

What it does

• Selects a product-label photo/image and reads ingredient text with in-browser OCR
• Lets you review and correct OCR text before ingredient matching
• Also accepts pasted INCI/English or Japanese ingredient text without a photo
• Organizes local-dictionary results as Dictionary match, Additional review, or Unmatched
• Filters results by state and lets you move through review/unmatched items
• Shows conservative spelling/OCR candidates for eligible unmatched entries and applies one only after an explicit user action

How to use

1. Choose From photo/image or From text. Photo mode is the default.
2. Choose the source-label type: INCI / English or Japanese label.
3. In photo mode, select an image, run OCR, and visually review/correct the recognized text.
4. In text mode, paste the ingredient label directly into the textarea.
5. Click Check ingredients to match the names.
6. Use result filters, previous/next review navigation, or an explicit unmatched-name candidate when useful.

What the result states mean

• Dictionary match: the entered spelling matched a maintained INCI name, Japanese name, registered alias, or other reviewed exact naming route.
• Additional review: a cue to check more official information or usage context. It is not a danger rating.
• Unmatched: the current dictionary could not identify the entered spelling. It does not mean the ingredient is unsafe.

The Japanese-label tab performs Japanese-name/alias dictionary matching. It is not a machine-translation feature.

OCR notes

• OCR can omit, split, or misread text depending on image quality, angle, glare, and text size.
• Always compare recognized text with the actual product label and correct it before relying on the match result.
• Near-match candidates are spelling/OCR repair hints only; they do not automatically identify an ingredient.
• Applying a candidate edits the input but does not automatically rerun the analysis. Run the check again when needed.

Limits

• FastScan does not assign SAFE / CAUTION / RISK safety ranks to ingredients or products.
• It cannot determine concentration, the complete formulation, individual suitability, allergies, pregnancy suitability, drug interactions, or regulatory status from the label alone.
• It does not provide medical/dermatological diagnosis or certify product safety.
• A known ingredient may still appear as Unmatched when its spelling is not represented in the maintained dictionary.

Data and network behavior

• Selected-image preview, OCR-text editing, ingredient matching, and result filtering are handled in the browser.
• The Tesseract.js OCR library is loaded from the external unpkg.com CDN, so the page is not fully offline.
• The tool does not provide saved scan-history or saved-result features.
• The UI language preference may be stored locally in the browser when available.

Lite vs FastScan

• FastScan: when you want to start from a product-label photo/OCR or use the more detailed review workflow
• Lite: when you already have the full ingredient text and want a faster paste-first workflow