Computer Vision for
Retail Digitization.

A camera-first onboarding tool that turned a retailer's shelves into a published storefront.

First Startup

Co-founded with Dr. Eli Osherovich, Senior Applied Scientist at Amazon (Alexa Shopping / Amazon Go), now leading AI & infrastructure at Google. Bootstrapped with no external capital.

~150
Products / hour
onboarding target
3
Computer Vision
generations shipped
60M+
Unique SKUs
database
The Problem

The Onboarding Bottleneck

Physical retailers need an online presence but lack the technical skills. Manual product upload is the single largest barrier to eCommerce adoption.

Skill gap

Brick-and-mortar retailers lack in-house digital teams to photograph, categorize, and list products.

Manual upload cost

~3 minutes per product × 50 products = 2.5 hours per store before a single photograph is taken.

No frictionless path

No existing solution generated a complete product catalog from the camera alone.

System Design

Technical Architecture

A camera-first onboarding tool that turned any retailer's shelves into a published storefront with all essential eCommerce features built-in. A Progressive Web App over a WooCommerce-vendor backend, with a two-database split and a vision pipeline.

PWA topology

Camera, push, and offline capabilities without an app-store dependency, updates rolled out the moment a seller refreshed.

  • RTL + Hebrew, Google Auto-Translate for the launch market
  • Push notifications, react-web-notification (orders, low-stock)
  • Analytics, Mixpanel events / GA4

Two-database model

One is shared across vendors, the other is private to each vendor.

  • Global Product Database, shared, canonical, retrieval-indexed (60M+ SKUs)
  • Store Product Database, vendor-scoped inventory, pricing, overrides
  • Integration, WC Marketplace REST API

Vision stack

The product replaced its Computer Vision front-end three times.

  • V1, ML Kit on-device barcode scanning + GPD lookup
  • V2, Vision-AI OCR + foundation-model background removal
  • V3, RAG retrieval over the GPD on video frames

Three Generations of Vision

Each version removed a layer of friction from the seller's upload workflow.

Version 1.0

Barcode → auto-upload

Barcode + Database
Version 1 flow Camera ML Kit barcode GPD lookup 60M+ SKUs Auto-fill WC publish
Scanner: Google ML Kit, on-device · zxing-cpp-emscripten fallback
Auto-fill: Global Product Database lookup per UPC / EAN
Inventory: bulk upload + per-product overrides
Target: store onboard + 30 products in < 1 hour
Version 2.0

Image recognition

Image + OCR
Version 2 flow Photo Vision OCR on-package text GPD enrich attributes · tags Vision cleanup bg removal WC
Input: single product photo from the camera
OCR: Vision AI text extraction from packaging
Generation: description, tags, categories, attributes
Cleanup: Foundation-model background removal
Target: store onboard + 50 products in < 1 hour
Version 3.0

Video + RAG over the GPD

Video + Retrieval
Version 3 flow Video pan across shelf Frame split N frames / sec RAG over GPD retrieval-grounded Draft batch multi-SKU WC
Input: video stream, frame-by-frame product recognition
Retrieval: dedicated RAG over Global Product Database
Attributes: enhanced extractor with synonym support
Fallback: single product photo from the camera
Target: store onboard + 150 products in < 1 hour
Traction

Traction & KPIs

Bootstrapped to $10K MRR within 6 months after 3 years of R&D.

~150
Products / hour onboarding target
$10K
MRR, from SMB customers (see Pivot)
6
Customers in production

Started like this:

Field testing the Seller App with end-users inside a retail location
Field testing · May 2022

End-user usability session

Daniel Manzela working with a retailer during in-store Seller App field testing
Field testing · May 2022

On-floor co-discovery

Ended like this:

Captured Session · Barcode Scanner

Barcode scanning workflow

Captured Session · Computer Vision

Background removal & image normalization

Pivot

Why We Pivoted

Human-in-the-loop

Retailers and their on-ground employees were not incentivized to perform the setup themselves. Onboarding stalled.

The discovery

Big-box retailers, operating eCom sites without local visibility, turned out to be the right buyer for a frictionless, automated, large-scale eCom-per-location setup.

The outcome

Pivoted from SMB self-service to enterprise frictionless solutions.

Retrospective

What I'd Do Differently

Despite building a technologically supreme product engineered by top minds from this field, our iterative feedback and hands-on high-touch approach revealed that brilliant tech isn't enough because "human-in-the-loop" was the single biggest problem and barrier to adoption.