From a mark to an estimate
- The browser trims surrounding whitespace, removes spaces, and converts letters to uppercase.
- The decoder checks the selected brand against its supported patterns and validates the date-bearing fields.
- It applies that format’s year-cycle and date-precision rules, then attaches the relevant explanation and source record.
Normalization does not repair an unreadable mark. The checker cannot know that O was intended to be 0, or that a product reference was entered instead of the manufacturing code. Photograph the original input before interpreting a result.
Confidence labels describe the date, not the bottle
| Label | Meaning |
|---|---|
| High | Reserved for a specific date supported by independently corroborated evidence without unresolved cycle ambiguity. No active format currently produces this label. |
| Estimated | A supported rule matches, but a repeating year, limited precision, or third-party evidence limits the conclusion. |
| Unknown | No approved matching rule or usable supported date was found for the selected brand and input. |
Some source records contain a rule-level confidence value such as high or medium. That metadata is separate from the result label. In particular, Dior’s 24-year assumption is now shown as Estimated. None of these labels grades a seller, the smell, or authenticity.
What is recorded for each rule?
The library stores the brand, supported structure, date-bearing positions, source names and URL, rule confidence, and a recorded verification date. Brand pages describe only the implemented subset. A named source and a working parser do not establish manufacturer endorsement or coverage of every product era.
| Brand | Recorded verification date |
|---|---|
| Chanel | 2026-08-29 |
| Christian Dior | 2026-08-29 |
| Yves Saint Laurent | 2026-08-29 |
| Jo Malone London | 2026-08-29 |
| Giorgio Armani | 2026-08-29 |
| Creed | 2026-08-29 |
| Versace | 2026-08-29 |
| Tom Ford | 2026-08-29 |
| Clinique | 2026-08-31 |
| Gucci | 2026-08-31 |
| Burberry | 2026-08-31 |
The main cited format source is the Scento format dataset. During this editorial update its endpoint could not be retrieved for independent rechecking, so the historical record dates above have not been advanced. Most records share this source; they should not be represented as multiple independent confirmations. Real bottle evidence and direct manufacturer clarification remain important limitations.
How repeating years are handled
- Chanel’s supported A–H scheme lists up to three recent matching years from an eight-year cycle, anchored in the implementation at A = 2008. It does not resolve the bottle to one year or extend that anchor to every vintage era.
- Dior’s supported letter scheme uses a 24-letter cycle anchored at A = 2000. The decoder selects the latest supported month/year candidate not later than the current month.
- Tom Ford and Jo Malone use a supported month-letter/year-digit subset. If the candidate month would be in the future, the decoder steps back a decade.
These are implementation policies, not proof of a bottle’s production period. The browser’s local date is used for cycle resolution. Brand-page worked examples have an explicit reference date so a future build does not silently change the meaning of an example. Consult the conflicting-date guide when comparing tools.
Inputs and the current product
The core checker is free and has no account requirement. Its current decoding runs in the browser using the embedded rule library; the form does not upload code photographs or submit the typed code to a decoding server. Loading the website still requires ordinary web requests. This describes the present implementation, not a promise about every browser extension or third-party service.
Save a result that can be reviewed
Keep the brand, exact typed code, original photos, output, confidence, source note, and check date together. Document what is observed separately from what is estimated. If the mark is Unknown, retain it instead of replacing it with a guessed date.
Copied codes, exchanged boxes, refills, damaged marks, formula changes, and storage history cannot be resolved from characters alone. Use the secondhand checklist, storage guide, and identifier guide for context.
Frequently asked questions
Is ScentDecoder an official brand tool?
No. It is an independent production-date reference using a local rule library. It is not affiliated with the fragrance brands or a manufacturer’s production database.
Why are supported results labeled Estimated?
The active formats involve repeating years, limited date precision, or third-party evidence. A pattern match cannot remove those limits. A stored rule confidence value is not a manufacturer-confirmed date.
Will you add my brand?
A brand needs a reviewable rule before it can produce a date. Until a suitable rule is available and implemented, an unsupported input remains Unknown.
Can the methodology change?
Yes. Better evidence may change a rule, its support scope, or its confidence. Save the code and check date, and revisit important estimates after an update.