Fit Receipt
A privacy-first virtual try-on AI agent product with a built-in judgment model called JEV. It generates a traceable AI decision receipt for each outfit recommendation, and user photos are processed entirely within the browser without leaving the device.
Timeline (last 90 days)
Fit Receipt ranked 19th with 73 votes in the Vercel Day event on Product Hunt.
Fit Receipt is a privacy-first virtual try-on agent whose core feature is virtually trying on lingerie in the browser, with each selection accompanied by an AI verdict receipt.
Fit Receipt is positioned across three categories: Productivity, Fashion, and Developer Tools.
JEV (TypeSafe) is the technical core of Fit Receipt — a judgment model that specializes in scoring trade-offs and providing calibrated confidence levels.
JEV is not triggered every time; the product is designed to know when to invoke JEV, and when the confidence of a judgment result is too low, the system will remain in a pending state rather than forcing an answer.
Fit Receipt's entire workflow can still run without JEV, and all judgments are recorded.
The division-of-responsibility principle Fit Receipt follows is: the Agent handles chat, while code protects facts and pricing.
Fit Receipt's photos remain local in the browser, the server stores no photos, the JEV model can only see structured fields and never the photos, and the API is rate-limited.
Fit Receipt is an open-source reference implementation, deployed on Vercel.
Fit Receipt received 11 comments and 73 upvotes on Product Hunt.
All Facts (8)
Fit Receipt is a privacy-first virtual try-on agent whose core feature is virtually trying on lingerie in the browser, with each selection accompanied by an AI verdict receipt.
50%UnverifiedFit Receipt is positioned across three categories: Productivity, Fashion, and Developer Tools.
50%UnverifiedJEV (TypeSafe) is the technical core of Fit Receipt — a judgment model that specializes in scoring trade-offs and providing calibrated confidence levels.
50%UnverifiedJEV is not triggered every time; the product is designed to know when to invoke JEV, and when the confidence of a judgment result is too low, the system will remain in a pending state rather than forcing an answer.
50%UnverifiedFit Receipt's entire workflow can still run without JEV, and all judgments are recorded.
50%UnverifiedThe division-of-responsibility principle Fit Receipt follows is: the Agent handles chat, while code protects facts and pricing.
50%UnverifiedFit Receipt's photos remain local in the browser, the server stores no photos, the JEV model can only see structured fields and never the photos, and the API is rate-limited.
50%UnverifiedFit Receipt is an open-source reference implementation, deployed on Vercel.
50%