AI-to-AI Commerce.
For real products.
ChatGPT, Claude, Perplexity and Gemini already recommend products to your customers — without mentioning you. CatyAI makes your catalog discoverable, trustworthy and transactable for them. No site rebuild, no crypto, no handing over money or customer data.
The merchant remains Merchant of Record. Payment and delivery stay fully with them.
Zero personal customer data sent to AI models. Compliance by design.
The catalog becomes visible to agents without any changes to the existing store.
Agentic commerce is already real — but on someone else's turf
Major protocols are already building infrastructure for agent-to-agent transactions: Stripe is working on agent-initiated payments and machine-readable catalogs; Coinbase and the x402 ecosystem launched markets where agents discover and pay for digital services in stablecoin; Google is pushing agent-to-agent payment protocols. The direction is no longer speculation.
All these platforms solve digital services commerce, settled in crypto, with the agent as a wallet-holding buyer. None address the traditional merchant — one who sells physical products, collects in local currency, has no x402 wallet and will not rebuild infrastructure. Exactly where CatyAI operates.
The problem: AI Blindness
LLM-oriented crawlers read raw HTML and ignore JavaScript executed in the browser. Stores built as dynamic apps appear to agents as nearly empty pages. The product exists, but for AI it's invisible.
Browser-rendered content never reaches the AI crawler. The product can't even be cited.
Even if a product is mentioned, the agent can't verify who the seller is, whether the price is real, or if the offer still exists.
Without a structured offer and a clear purchase path, the recommendation stays a good word, not a transaction.
What is CatyAI
CatyAI is a discovery and attribution layer for agentic commerce. It sits between the AI agent and the merchant: makes the catalog discoverable, validates offers deterministically before any recommendation, and directs the customer, via redirect, straight to the merchant.
- — Discovery layer for agents
- — Attribution layer (who brought the order)
- — Deterministic offer validation
- — Bridge between AI and the real merchant
- — Not Merchant of Record
- — Does not process payments, does not touch money
- — Does not collect personal customer data
- — Does not use crypto / wallets / x402
How it works, step by step
Signed catalog
The merchant's catalog is indexed semantically and signed cryptographically (Ed25519), together with verified identity data (NAP). The agent doesn't receive free text it must believe, but data whose authenticity it can verify mathematically.
Discovery
When an agent searches for a product, it queries the semantic catalog directly, in a machine-designed format — not via page scraping. The response is structured, complete and instant.
Deterministic validationAPO → CEE → TSO
Before an offer is recommended, it passes through a deterministic chain: the signature is verified, the real existence of the offer and the price correctness are checked. Never an unsafe or invented offer.
Redirect via shop.catyai.io
The link the agent recommends lives on the CatyAI domain — shop.catyai.io/{store}/{product} — which makes a 302 redirect to the real product page on the merchant's site, carrying a reference parameter. The customer buys and pays at the merchant, as usual.
Attribution
The completed order is correctly attributed to the discovery that generated it, through the ref=caty_… parameter carried from the shop.catyai.io redirect. The merchant sees exactly which sales came through the agentic channel.
The trust architecture
In a world where agents recommend automatically, trust cannot be a marketing promise — it must be verifiable. CatyAI builds it from three pieces.
Identity
The merchant's name, address and contact are verified and tied to the catalog. The agent knows who it's talking to.
Integrity
Every data set is cryptographically signed. Any unauthorized modification invalidates the signature — data cannot be falsified in transit.
Correctness
An offer reaches the agent only after deterministic validation. Never an unsafe price or a presumed availability.
Classic SEO vs. CatyAI
| Capability | Classic SEO | CatyAI |
|---|---|---|
| Cited by agents (GEO) | Partial | Yes |
| Trusted (NAP + signature) | No | Yes |
| Transactable (validated offer) | No | Yes |
| Attribution on result | No | Yes |
Where CatyAI fits
CatyAI doesn't compete with agentic payment rails — it complements them. Global protocols solve agent-to-agent payment for digital services. CatyAI solves the last mile for physical products of real merchants, who don't have the time and budget to rebuild their infrastructure.
- — Sell digital services / APIs
- — Buyer: agent with wallet, no human
- — Settlement in stablecoin / on-chain
- — Require merchant API and integration
- — Physical retail products (RO / EU)
- — Buyer: human finalizes at merchant
- — Classic payment, in local currency, at merchant
- — Zero site rebuild, zero mandatory API
For merchants
- ●Agentic visibility without rebuilding the site — the catalog becomes discoverable as-is.
- ●You stay Merchant of Record — payment and delivery stay fully with you.
- ●Zero personal data of customers sent to AI models.
- ●Commission on result — you pay only for confirmed, delivered and collected orders.
You pay when you sell. Not before.
A percentage of the value of orders brought through the agentic channel, applied only to completed orders. No startup subscription to be discoverable.
I want to be discoverableFrequently asked questions
A discovery layer for AI-to-AI commerce: agents (ChatGPT, Claude, Perplexity, Gemini) discover, evaluate and recommend products from real merchants, reading a cryptographically signed catalog instead of heavy web pages.
Make your products discoverable for the next generation of buyers.
Connect your store and enter the agentic catalog. No site rebuild, no risk.