How Dr. Vin AI car inspection works: upload used car listing photos, get instant AI damage detection with component grades for exterior, interior, tires, wheels, and glass. Each finding includes repair cost estimates and confidence scores. Pre-purchase vehicle inspection from photos without visiting a mechanic.
How Dr. Vin Works
You upload photos. Dr. Vin tells you what's actually wrong with the car and what it'll cost to fix. Here's what happens in between.
You bring the photos
Listing screenshots, your own shots, dealer photos. 3 to 30 images from any source. That's all Dr. Vin needs.
Each photo is indexed, resized for optimal model input, and stripped of EXIF metadata. If a license plate is visible, Dr. Vin reads it, resolves it to a VIN through NHTSA's vehicle database, pulls year, make, model, and trim, then redacts the plate from storage. The vehicle identity is locked in before analysis starts.
Dr. Vin looks at everything
Every photo is analyzed for scratches, dents, paint issues, tire wear, interior damage, and signs of prior accidents. Each finding gets a confidence score and a repair cost estimate.
Dr. Vin runs a multi-call ensemble across multiple vision models. Each photo is scored against a per-component rubric spanning hundreds of individual condition variables. Findings are cross-referenced across images, and observations only survive if a majority of calls agree, using Jaccard similarity matching to align detections across runs.
Repair costs pull from a proprietary cost matrix sourced from CCC, Mitchell, Manheim, and four other industry databases, then adjusted for regional labor rates across seven U.S. zones. Dr. Vin uses the Bernoulli equation to weight each finding by its confidence score. Value impact multipliers translate raw repair cost into what the damage actually means for the car's worth. P2P discount factors account for the private-party pricing gap. Condition impact ranges are derived through variance propagation across all observations, not averages or guesses.
None of this would be possible without Phototology's Aletheia Engine, turning photos worth a thousand words into a thousand words.
You get a report, not a data dump
Overall condition grade. Component scores. Every finding with photos, locations, and what it costs. A Walk-In Brief that tells you whether the listing earns your Saturday. Negotiation leverage. All of it organized so it actually makes sense.
The report is structured, not generated. Findings are grouped by severity and component, color-coded by condition zone, and presented as a narrative that tells the story of the car's condition. The goal is not to impress you with data. It's to make the next conversation with the seller a short one.
What Dr. Vin actually sees
Real photos. Real findings. Drag the slider to see what Dr. Vin found on this 2004 Lexus LS 430.
front bumper · driver side fender
hood-to-fender
driver door · rear quarter panel
passenger side · rear bumper · trunk lid
both headlights
Every finding is spatially located on the original photo using bounding box coordinates. The confidence score is calibrated across the model ensemble. When multiple models agree on a finding, confidence goes up. When they disagree, Dr. Vin flags it and tells you.
Honest about what we are
Dr. Vin is a photo-based screening tool. He's really good at what he does. He doesn't pretend to do what he can't.
What AI sees
- Paint defects and color mismatch
- Body damage, dents, and panel gaps
- Tire condition and tread wear
- Interior wear and damage
- Glass and light damage
- Accident and flood indicators
What requires hands-on
- Engine and transmission
- Brake system
- Hidden frame damage
- Electrical and HVAC
- Suspension and drivetrain
- Odometer verification
That's why we built the 60-point inspection checklist for everything you need to check in person.
That's it. Photos in, answers out.
60 seconds from upload to report.
Run ReportFree instant grade. Full report $15.