Ingest
POS · ERP
Priority, Comax, SAP, Shopify — SFTP, CSV, or API, every 15 minutes.
EasyToBuy.ai
An online store doesn't make you a tech-driven retailer. Meet the silent engine powering your fastest-growing competitors — and orchestrate your own growth.
Where retailers actually can grow.
The brain behind the brands eating your market share.
The silent engine powering your fastest-growing competitors.
The operating system for retailers ready to scale.
The operating system for the next retail market leaders.
The operating system for retailers destined to lead.
The ultimate moat for mid-market retail.
An online store doesn't make you a tech-driven retailer. Orchestrate your growth.
80%
Of retail spend still happens in physical stores
40–60%
Of inventory records are wrong at any time
15 min
End-to-end feed freshness, store to Google
1B+
SKU updates per minute pipeline target
How the Router Works
No drawn-out integrations, no marketing feed plugins pretending to be infrastructure. Five stages, purpose-built in Rust, run continuously.
POS · ERP
Priority, Comax, SAP, Shopify — SFTP, CSV, or API, every 15 minutes.
ETBCS
Map every raw feed into our canonical standard — SKU, GTIN, variants, state.
AI Tier
Tiered cache → GTIN lookup → self-hosted Llama → paid API. Lowest-cost path first.
Confidence
Green: publish. Yellow: resolve. Red: suppress & flag. Phantom inventory, killed.
Surfaces
Google Merchant Center, Meta, TikTok Shop — and the AI agents that query retail.
The Canonical Standard
ETBCS — the EasyToBuy Canonical Standard — is a universal product model grounded in GS1, Google Merchant Center, and real retail lifecycle research. Parents, variants, aliases, and inventory state — all normalized.
product.etbcs.jsonl
{
canonical_sku: "FOX-TSHIRT-BASIC-2026-WHT-M",
gtin: "7290001234567",
item_group_id: "FOX-TSHIRT-BASIC-2026",
variant_attrs: { size: "M", color: "White" },
price: { regular: 99.90, sale: 79.90, currency: "ILS" },
inventory: {
qty: 8, atp: 5, buffer: 3,
updated_at: "2026-04-14T09:22:03Z"
},
aliases: [
{ tenant: "fox", source_sku: "10042" },
{ tenant: "mega", source_sku: "CHRG-AP" }
],
confidence: "green" // publish to GMC
}Confidence Scoring
If we wouldn't bet on a record, neither will Google. Every SKU enters one of three lanes. A buffer threshold (< 3 units → out-of-stock) protects shoppers and retailers from the oldest retail lie: “we have it.”
Green
≈ 80%
GTIN matched, variants resolved, inventory state within buffer. These flow straight to Merchant Center, Meta, and the Global API without human review.
Amber
≈ 15%
Name drift, missing GTIN, ambiguous variant. Routed to NLU, then the paid LLM tier only if needed. Published once confidence crosses the line.
Red
≈ 5%
Records like “1” or “general” or truncated Hebrew. We never publish these. We report them back to the store owner with a cleanup suggestion.
The Platform Ahead
We ship what exists and label what doesn't. The Data Router runs in production today. The rest are on the roadmap with their quarters attached — not “coming soon.”
Intelligence layer that crosses search intent with real-time inventory. Demand predictions, dead-stock routing, missing-money maps.
When the nearest store is out, intercept the click: pickup from another branch, ship-to-store, or same-day courier. White-labeled under retailer domains.
Hebrew-first RAG discovery. Geo-aware inventory ranking. Shabbat-aware hours. The keystone service for first-party product search.
Real-time physical inventory for AI agents — OpenAI function calls, Anthropic MCP, Gemini. RAG-only licensing; no training rights.
Design Partner
“A tier-1 global sports retail franchisee — part of the Fox Group / Harel Wizel network — is running the Router pilot on two anchor stores with a hundred high-velocity SKUs.”
Ecosystem & Integrations
Built for the Israeli Retail Stack
We're building EasyToBuy from Tel Aviv, for the retail stack we know best — then exporting the pattern. Hebrew, RTL, and local infrastructure are first-class, not retrofits.
Ready?
Two anchor stores. A hundred SKUs. A fifteen-minute feedback loop from shelf to Google. That's how we start together.