01Check trade-ins before you give credit.
A client brings a piece back for a voucher. Your advisor checks it in store, before the credit is issued. When the model allows it, the report gives an estimated era.
Stores · Trade-in · Clienteling
No return, trade-in, delivery or listing without a check.
Authenticity checks for fashion and resale. Your team takes a few photos. LegitV checks them against a checklist built for that exact model, and tells you what it found.
Buying second‑hand yourself? Check one item →
- 5 to 34points checked, depending on the model
- ~15sAPI pre-check (< 2 min at the desk, up to 100× faster)
- From €0.65per check at API volume (€1.67 in self-serve packs)
59 model checklists ready today · your models nextSneakers · Bags · Denim · Streetwear
- Photo readingGoogle Gemini3.8 Flash · every checklist point in one pass
- Listing triageTypeSafe JevReplica wording, price, provenance
- ReportSHA-256Fingerprint of the photos and the verdict, checkable through the public API
- PhotosParisStored on AWS, eu-west-3 region
Four ways a fake gets in. A check for each.
One engine, set up for each step of your process. These are examples. We fit the checks to yours.
01
02Check returns before you refund.
A fake swapped into a return costs you the refund, then goes back on the shelf. Your returns team checks it at the desk, in minutes.
Returns · Customer service
03Check every item at the door.
Buy-back, consignment, second-hand programmes. Nothing enters your stock without a check.
Logistics · Buying
04Screen listings before they go live.
Marketplaces and vintage resellers send photos through the API. Clear cases go live. Doubtful ones go to your experts.
Marketplaces · Resellers05
Your step here.
After-sales, repairs, production review, insurance claims. Tell us where you need a check. We build it around your process.
Tell us your process →You draw the line.
Every check ends on a score, from likely fake to likely authentic. Above your line, the item clears. Under the other, it is held. In between, an expert decides, with every point already laid out.
Decided without an expert
75%
45 of 60 checks. 15 go to an expert.- 6HoldLikely fake, under 40
- 15ExpertBetween 40 and 80, every point laid out
- 39AcceptLikely authentic, 80 or more
Most returns are genuine: clear them fast, look closely at the doubtful ones.
Sample batch and example lines. You set your own. Your experts keep the last word.
Your spec sheet becomes the checklist.
We start from your product data. Each part of the technical sheet becomes points to check, with its weight and the photo it needs.
Technical sheet · Ref. 0000 · Season —Checklist · 7 points
- Cover pageReference, season, colour, factoryKeyModel, season and colourPhoto · Main label
- Technical sketchesFront, back, detailsMajorSilhouette and placementPhoto · Front and back
- Bill of materialsFabrics, hardware, suppliersKeyEngraved hardware, zip, buttonsPhoto · Hardware close-up
- LabellingMain label, composition, made-in, sizeKeyLabel type, made-in for that seasonPhoto · Care label
- ConstructionSeams, stitches per cm, finishesMajorStitch density and finishesPhoto · Signature seam
- MeasurementsChart per sizeMinorKey measurements, when in doubtPhoto · Flat, with a tape
- ColoursColour referencesMinorShadePhoto · Daylight, front
Your whole range, or a few models to start.
- Guided photos.
The app asks for the photos each model needs, point by point. No guessing what to shoot.
- Data at the core.
Each point carries its weight, the zone to look at and reference photos. Expert feedback flows back into the checklists.
- Built with your experts.
We write the checklists from your sheets and your experts’ know-how. A pilot on a few models first, then the range.
An AI built for one job.
Not a chatbot guessing from a photo. A pipeline made for authentication. Every check runs the same three steps.
Read every photo
Finds the labels, codes, stitching and hardware. Notes what each photo shows.
- Photo 1Side view
- Tongue labelFound
- Size tag codeRead
- Heel tabBlurred
Find the references
Looks up what an authentic one looks like, for this model and this code.
- Label fontReference
- Code formatValid
- Production eraMatched
- Known fakesCompared
Check every point
Each point is a match, not seen, or a red flag. Key points weigh more.
- Tongue labelMatch
- Size tagMatch
- StitchingMatch
- Heel tabNot seen
Likely authenticReport ready · every point explained
Sample run. A point we cannot see is never counted as a match.
The Point-by-Point Method · Why pre-checking changes the math
Not replacing your experts. Multiplying them.
Sending every piece to a physical hub takes days and drowns senior authenticators in routine checks. Asking a general chatbot produces confident guesses on blurry photos. LegitV pre-checks every named point upstream so your experts only judge the borderline 15–30%, with every crop and reference already side by side.
- Up to 100×faster upstream pre-check
~15s via API (or under 2 min with photos at the counter) vs. 15–25 min of manual reference lookup and inspection.
- 10×–30×lower cost per piece checked
From €0.65 per check at API volume (€1.67 in self-serve packs) vs. €15–€50+ to route every item through a physical hub.
- 3×–5×more pieces per expert, higher accuracy
~70% of clear cases settle upstream. Experts decide the remaining 15–30% in 30 seconds without visual fatigue.
| Dimension | 01 · 100% Manual (no pre-check) | 02 · General LLM / Vision chatbot | 03 · LegitV Pre-Check + Expert |
|---|---|---|---|
| Speed per piece | 15–25 min (or 2–5 days shipped to a hub) | 15–30s of unstructured prose | ~15s API / < 2 min at the desk |
| Cost per check | €15–€50+ in hub handling & shipping | Low API cost, high cost on the first missed fake | From €0.65 at API volume (€1.67 in packs) |
| Blurry or hidden label | Manual back-and-forth to request photos | Guesses from context with high confidence | Marked Not seen — never counted as a match |
| 1 bad key point + 9 clean seams | Risk of eye fatigue after 200 items/day | Averaged into a flattering ~85% score | Immediate Key (×3) veto → Hold / Likely fake |
| Role of your human expert | Spends 70% of the day on obvious items | Re-checks everything from scratch | Decides only the 15–30% borderline cases, every point laid out |
Sources: LegitV production API (~15s) & credit packs (€0.65–€2.99) · OECD/EUIPO (2025): 62% of seized fakes are apparel, footwear & leather goods, 79% in small parcels (<10 items) ·NRF (2024): $101B–$103B in retail return fraud · FINER & HAC VLM studies (2024–2025): uncalibrated VLMs output 90–95% confidence on fine-grained visual errors.
One red flag can be enough.
Key points weigh more than details. Try it: change a result and watch the verdict.
Tongue labelKeyfont, spacing, ® mark
Inner size tagKeystyle code, dates
Box labelMajorcode matches the shoe
StitchingMajordensity, straight lines
Heel embroideryMinorheight, thread
OutsoleMinortread pattern
Checklist seenSimplified
92%
Key ×3Major ×2Minor ×1
Likely authentic
Everything we can see matches.
Simplified. Each model has its own checklist, up to 34 points.
Every part. Every stitch.
LegitV looks at the parts fakes get wrong. Bags, denim and your own pieces get their own parts list.
- 1LacesWeave, aglets, length
- 2Tongue tagFont, spacing, ® mark
- 3UpperPanels, stitch density, leather grain
- 4InsolePrint, glue marks underneath
- 5CupsoleShape, stitch line around the toe
- 6OutsoleTread pattern, mould details
For your teams. For your clients. In your systems.
An internal tool, a service you offer, or an API behind your listings. Often all three.
- 01 · Your teamsA web app, nothing to install.
On a phone or a tablet. At the returns desk, in the stockroom, in store.
- 02 · Your clientsA service you offer.
Your advisor runs the check with the client and hands over the report.
- 03 · Your systemsAn API.
Send photos or a listing link. Get the verdict and every point as JSON.
Three steps. Then you know.
No expert to find. No days of waiting. A verdict in minutes.
- Step 1
Pick the model.
Search the brand and the model. Each one has its own checklist.
Step 2Add the photos.
The app asks for the shots each point needs. Taken on the spot, or from a listing link.
- Step 3
Read the verdict.
Likely authentic, likely fake, or more photos needed. Every point is explained.
Playbooks Returns · Trade-ins · Product sheets · Stock intake · Resale
Where fakes get in, and how teams check.
Returns, trade-ins, product data, quality control and resale: one playbook per step, backed by industry studies and research.
Stopping Counterfeit Return Fraud and Wardrobing Before the Refund Clears
How luxury houses, multi-brand retailers, and e-commerce warehouses stop replica swaps at the returns desk in under two minutes.
Retail Ops · E-Commerce · Returns WarehousesRead playbook
Turning Your Technical Sheets into Living Authentication Checklists
How a maison’s product specs and expert know-how become a repeatable, point-by-point inspection grid for every season.
Maison Heritage · Merchandising · Product DevelopmentRead playbook
Why General-Purpose Chatbots Fail at Authentication—and How Point-by-Point Pre-Checks Multiply Your Experts
Why we took the opposite direction from whole-image chatbots: never count a blurry detail as a match, let one key red flag veto the piece, and pre-check in seconds so experts focus on the 15–30% borderline cases.
CTOs · Marketplace Trust & Safety · Operations LeadersRead playbook
Checking Incoming Stock and Pre-Shipment Lots Against Model Checklists
How buying teams, consignment warehouses, and production controllers check every piece at the door with a phone or tablet.
Logistics · Buying Teams · Quality ControlRead playbook
Running In-Store Trade-Ins and Pre-Listing Checks with Shareable Proof
How boutiques, vintage resellers, and marketplaces verify second-hand pieces before issuing credit or publishing a listing—and prepare for EU ESPR traceability.
Clienteling · Vintage Resellers · MarketplacesRead playbook
Pay per check. Or a plan built for your volume.
One credit is one check. If a check fails on our side, the credit comes back.
BusinessBrands · Stores · Resale
Tailored
Per check, or a monthly plan per store with checks included.
- Checklists set up for your references
- A pilot on a few models first
- Web app for your teams, API for your systems
- A lower price per check as volume grows (from €0.65 via API)
- Set up with your experts
Pay per checkNo subscription
€2.99
a check. Down to €1.67 in packs.
- 1 check€2.99€2.99 each
- 11 checks€24.99€2.27 each
- 28 checks€59.99€2.14 each
- 60 checks€99.99€1.67 each
Personal
Buying second‑hand yourself?
Check one item before you pay. Same engine, same checklists. On Vinted, Vestiaire Collective or eBay.
Pay per checkA verdict in minutesLegitV for individuals →
Questions. Short answers.
What does LegitV do?
LegitV checks if a fashion item is authentic from its photos. It compares them to a checklist built for that exact model and gives a verdict with every point explained.
Can you set it up for our products?
Yes. We build a checklist for each reference you want covered, from your technical sheets and with your experts: your whole range or a selection. A pilot on two or three models is a good start.
Does it replace our experts?
No. It pre-checks every point in seconds (~15s via API or under 2 minutes with photos at the counter), settles the ~70% clear cases upstream, and sends the 15–30% borderline ones to your experts with every crop and reference already laid out. LegitV gives a verdict and its reasons, never an automatic certificate.
How is it different from asking a chatbot about a photo?
A chatbot looks at the whole picture and answers with confidence, even when the label it needs is blurry. LegitV checks named points for that exact model, one by one: each one matches, raises a red flag, or is marked not seen. A point it cannot see is never counted as a match.
How reliable is it?
Each point has a weight (Key ×3, Major ×2, Minor ×1), and a red flag on a key point is enough to call a fake. The report gives a confidence level. When the photos are not enough, it says so.
What if a label is cut or a stamp is worn?
That point is marked not seen, not failed. The verdict rests on the points that are visible, and when too few are, the report asks for more photos or sends the item to an expert.
How does our team use it?
In a web app on a phone or a tablet, nothing to install. The app asks for the photos each model needs. Your systems can call the API instead.
How much does it cost?
For businesses, per check (down to €0.65 at API volume) or a monthly plan per store with checks included, sized to your volume. Individuals pay per check (€1.67 to €2.99), with no subscription.