# VAppraise — An appraisal you can explain

Team Lowball Protocol · Cox Automotive Hackathon 2026

VAppraise brings the appraisal workspace into a conversation with Claude. Start with a VIN and a few photos, work through condition, and see what drives the number.

## Animated hero: complete transcript

This 40-second concept animation uses a fictional vehicle, illustrative images and sample values. It does not create an appraisal or make a live offer. Pause, replay, reduced-motion support and a readable transcript are available on the web page.

**Dealer:** I’ve got a trade-in I need to value.

**Claude:** Let’s take a look. Snap the VIN and a few shots of the car. I’ll help you work through the appraisal.

**Appraisal workspace:** The dealer captures the VIN, identifies a fictional silver 2021 Performance Wagon and uploads three photos, including a close-up of a scraped wheel.

**Claude:** I can see curb rash on the wheel in your photo. Are the other wheels affected too?

**Dealer:** Yes, all four. It has 38,000 miles and needs tires, too.

**Claude:** I’ll include all four wheels in recon, plus tires and finishing. Here’s the impact.

Illustrative market context: $61,000. Wheels, tires and finishing: −$2,800. Illustrative retail estimate: $58,200. History is unverified; no history adjustment is assumed. These are sample values, not an offer.

## Companion 3D vAuto lot view

The lot view lives in the existing vAuto non-production MCP and includes hackathon output. It displays groups of vehicles in a 3D lot, with controls for days in inventory, ProfitTime metal, price minus cost and day supply. In the actual tool, users can orbit, zoom, select signs and inspect vehicles. The landing-page image is a static, AI-edited screenshot with dealer identity removed; it is not an interactive lot.

## The film

A dealer and customer smile and dance beside a silver SUV outside a dealership. This is a still from the team’s AI-generated hackathon film. The final branded cut is 2 minutes 59.6 seconds, with the VAppraise logo and AI-tool credits.

[Watch the final film in team Slack](https://coxauto.slack.com/archives/C0C1V2A8R3J/p1790555609441729?thread_ts=1790538171.120049) (team access required).

## AI tools used across the project and media

Claude; ChatGPT / Codex; Gemini; ElevenLabs; Suno; AWS DevOps Agent; Tesseract OCR.

## Further reading

[Project documentation](https://pages.ghe.coxautoinc.com/Hackathon-2026/lowball-protocol/) requires Cox enterprise sign-in. It documents the server’s recommended appraisal value, maximum approved buy, evidence states, and guarded non-production write flow. The simplified concept animation is not a substitute for that specification.