Hi Product Hunt — I built Pairvu because a product image can look polished while quietly changing the product itself. Pairvu adds a reference-based Visual QA step between image creation and publishing. You upload an approved original and the final generated, edited, or candidate image. Pairvu checks visible product identity details such as logo and label text, color, quantity, major components, and packaging shape, then returns PASS, REVIEW, or FAIL. The REVIEW state matters as much as PASS or FAIL. If a label is unreadable, a logo is hidden, or the viewpoint does not expose the same product face, Pairvu keeps that uncertainty visible instead of treating missing evidence as a match. Pairvu does not generate images, certify marketplace compliance, or replace human approval. It is a pre-publish product-fidelity checkpoint for ecommerce sellers, brands, agencies, and creative operations teams. The public beta includes a free plan, comparison examples, and a controlled Visual QA benchmark. I would especially value feedback on the evidence shown with each verdict and where this checkpoint would fit in your current product-image workflow.

Pairvu

Visual QA for AI product photography

Visual QA for AI product photography