Serum · SKU → country version → compatible reference
Box, label, batch, barcode, pump, cap, seal, typography and bottom mold.
Open beauty lab ↗EVIDENCE-LED PRODUCT AUTHENTICATION
We begin with the exact beauty product and version, obtain usable physical evidence, compare with a qualified reference, and state the limits of each conclusion.
Try the four-direction 360° sample ↗Synthetic experience preview · No live authentication is performed here
TWO CORE INDUSTRIES / DEMONSTRATION
Box · label · lot · seal · pump · cap · country and packaging variation
Explore beauty authentication ↗ 02 / FASHION & ACCESSORIESGeometry · leather · stitch · edge paint · hardware · serial · 360° detail
Explore fashion authentication ↗Initial focus: brands in Korea, Japan, Taiwan and China; physical evidence capability in China. Service availability depends on specimen and reference quality.
01 / FROM PRODUCT TO REPORT
A visual assessment begins with a particular physical specimen. Capture and comparison are inputs; findings, expert review and a report are separate stages.
CORE INDUSTRY / BEAUTY & COSMETICS
A bottle that looks similar can belong to another SKU, market or production version. Beauty assessment needs a matching reference, known authentic variation and physical context alongside visual comparison.
Initial product focus · An individual assessment still depends on the exact reference and evidence available.
AUTHENTICATION LABS / DEMO
A cosmetic batch and a leather bag require different photographs and known authentic variations.
Box, label, batch, barcode, pump, cap, seal, typography and bottom mold.
Open beauty lab ↗Front, rear, bottom, interior, leather, stitch, edge paint, hardware, engraving and serial.
Open fashion lab ↗01 / INTERACTIVE DEMONSTRATION
A synthetic example shows how a reference and a candidate can be compared feature by feature. Select an observation to inspect its evidence.
02 / INSPECTION TOOLS
Explore two focused workspaces. The viewer handles a sequence of product angles; the map records where a shipment was said to originate. Neither creates an authenticity conclusion.
Try the three-height synthetic sample immediately, or load your own matching photo sets. Drag in four directions and pinch to zoom.
Open workspace ↗ TOOL / 02Place a marker or record an address without coordinates. Preserve the declared source and the limits of location evidence.
Open workspace ↗03 / THE METHOD
The process starts with a real, identifiable specimen. The system organizes its evidence; analysis and expert review then address only what that evidence can support.
Record the product, variant and individual physical specimen. A seller's claim is context, not proof.
Submit whole-object and detail views. Check focus, coverage and missing angles before comparison.
Match the right product variant to qualified reference evidence. Document visible similarities and differences.
Separate AI-assisted findings, preliminary visual risk and any expert opinion. Preserve the limits of the result.
People obtain and review.The system preserves and connects.AI assists with visual differences.
MARKETPLACE → PHYSICAL SAMPLE / ILLUSTRATIVE TIMELINE
DEMO ONLY · Production events require a timestamp, source, status and linked evidence.
PHYSICAL EVIDENCE ROUTE
When a brand cannot obtain the suspected item directly, a controlled sample purchase can establish what object was actually inspected. Source records and custody must accompany the images.
Record the listing, seller, price and purchase source before acquiring a sample.
Keep packaging and delivery context with the identifiable object and its photographs.
Use eligible references, evidence-linked findings and expert review within a stated claim limit.
This is a service workflow description. The public preview does not place an order or create a case.
Explore interactive intelligence dashboard ↗
COUNTERFEIT INTELLIGENCE / DEMO · ILLUSTRATIVE DATA
Historical seller context does not decide the authenticity of a new product.
Marketplace · store · first/last seen · acquired samples · cases · related SKU · reappearance
A repeated physical pattern across qualified samples.
Batch printing changes; pump structure changes; bottom mold remains distinct.
Observe feature changes over time.
New and repeat sellers · samples · families · versions · affected SKU · marketplace
Illustrative concepts only.
04 / DECISION STANDARD
Similarity is not authenticity. Before any assessment, the system must establish whether the comparison is valid and the evidence is sufficient. Expert review remains a separate decision.
Confirm the source, product identity, variant and normal manufacturing variation of the comparison material.
Check the required views, focus and coverage. Ask for additional images when a significant feature cannot be inspected.
Keep AI observations, preliminary visual risk and a reviewed final opinion as distinct records.
05 / PRODUCT SUPPORT
Select a category to see a likely evidence route. Actual eligibility depends on the exact model, reference availability and the physical question being asked.
Synthetic, unbranded product visuals · Category examples, not eligible references
EVIDENCE GRAPH / CONCEPT
Concept preview; IDs are illustrative.
06 / RESULT EXPERIENCE
A result should explain what was inspected, what differed, what remains uncertain and who reviewed it. This preview illustrates the report structure.
Synthetic example · No authenticity opinion issued
When evidence or the reference is insufficient, the right result is to request more evidence or withhold a conclusion.
REPORT CENTER / FUTURE OUTPUTS
SYSTEM = source of truth · GPT = language and structure.
View the existing sample report ↗BRAND USE CASES
For beauty brands, a traceable record can connect product identity, marketplace context, acquired samples and evidence-linked reports. Each use still depends on qualified case evidence.
Compare the correct SKU and packaging version with an eligible reference.
Connect seller, listing and purchase records to an identifiable physical sample.
State findings, uncertainty and the review scope for brand or counsel use.
Repeated, qualified cases could later reveal trends by SKU, seller, marketplace, region and price. This preview does not present live intelligence or validated statistics.
GoodsInspect