Was it really broken, or just generated?
Some buyers now fake the cracked vase or the bent card with AI to get a refund. Fraydar examines the photo, compares it with your listing and drafts the appeal for eBay, Etsy or Vinted.
- Be first in line
- One email, when we open
- No spam
- Running image forensics0%
- Reverse-searching damage photosQueued
- Parsing photo metadataQueued
- Comparing listing photosQueued
- Drafting platform appealsQueued
Forward the buyer’s ‘it arrived broken’ photo and get a forensic report and appeal pack.
Some buyers now generate the torn sheet or the cracked vase with AI to get a refund without a return. Fraydar checks the photo for edits, metadata, lighting and reverse-image hits, compares it with your own listing photos, and writes the appeal. It sells evidence and reasons, never a bare verdict.
- Step 01
Forward the damage photo
Send us the buyer’s ‘it arrived broken’ photo, along with your own listing photos of the item.
- Step 02
We examine it like a lab
An ensemble of image forensics looks for edits and AI generation, reads the metadata, tests the lighting and runs a reverse image search.
- Step 03
You get the evidence and the appeal
A report that gives the reason for each finding, and an appeal letter written to the platform’s own policy, with the evidence attached.
Built for resellers and small shops selling high-value items.
A per-claim forensic report
How likely the photo is AI-made or edited, with the reasons, metadata and lighting checks, and reverse-image hits.
A side-by-side with your listing
The buyer’s photo set against your own listing photos, so differences are shown, not guessed.
An appeal letter for each platform
Written to eBay, Etsy, Vinted or Shopify policy, with the evidence attached.
First up: We start with a free single-photo check, shared in the seller forums where these refund threads already run.
The world just changed.
What changed, and where it was reported.
The OECD AI incident monitor logged AI-generated photos of damaged goods used for refund fraud on Vinted, Amazon and Fnac.
OECD.AIeBay trading-card sellers reported a spike in doctored damage photos. Forter says AI-manipulated image claims rose over 15% worldwide in a year.
ShopifreaksRetailers face a surge in AI-generated fake damage claims, and fraud rings now sell ‘returns-as-a-service’.
Retail Gazette
One person and 2 AI agents.
The agents, all day
- Running image forensics
- Reverse-searching damage photos
- Parsing photo metadata
- Comparing listing photos
- Drafting platform appeals
What a person does
No person handles each claim. Before launch, a person reviews how ‘likely fake’ findings are worded and checks the detector’s error rates, so a report never claims more than its evidence shows.
Can you prove a photo is fake?
No detector can. AI-image detectors still miss many fakes, so we give evidence and reasons you can check, and we say plainly when a photo looks genuine.
What if the buyer is honest?
Then the report should say so. Wrongly accusing a real customer is the worst outcome, which is why our wording is reviewed and every finding shows its reasons.
Is this for card chargebacks?
No. Fraydar is about the photo: whether the damage shown in an image is real. Card chargeback disputes are a different problem with their own tools.
Be first through Fraydar’s door.
- Be first in line
- One email, when we open
- No spam