The internet never got a commerce layer. We’re building an open one.
Intents Protocol is neutral, open infrastructure for commerce, built for a web where agents shop alongside people.
It’s the foundation for an honest, transparent future of commerce, and a sustainable new way to fund the open web.
Data, email and the web run on open protocols. Commerce runs on platforms.
The internet’s core protocols belong to everyone, and that’s why the web took off. Commerce never got that. What people see, what they’re nudged to buy and who gets paid all happen inside a few platforms, each with its own catalog, its own rules and its own interests.
The same product is described a hundred different ways, and every platform keeps its own version of the truth.
Ads, commissions, store brands: someone always pays for what shoppers see first, and no one can check how it was decided.
Most of what sellers spend to reach shoppers, and everything learned along the way, ends up with a handful of giants.
Now agents are starting to shop for people. Before they inherit the old rules, commerce needs a neutral layer.
A neutral commerce layer for the agentic web.
Intents Protocol gives people, agents and sellers common ground: an open record of every product, matching no money can buy, and a transparent way to share the value. It’s the infrastructure the agentic web is missing, and the next chapter of commerce.
Honest, transparent commerce
Products win on fit, under rules anyone can check. Trust isn’t promised; it’s built in.
Infrastructure for agents
One shared catalog and one set of rules, so any agent can find, compare and buy from any seller.
Funding for user-owned AI
Assistants earn from the commerce they help with, and the money never touches what they recommend. That’s how AI that answers only to its user can pay its own way.
A fair market for every seller
The corner store and the giant play by the same rules, and the best offer wins.
Built to be owned by no one
Like email and the web, it’s built so that no company controls it. Not even us.
Anyone can build
Assistants, apps and tools of every kind plug into the same graph and rails, and earn from day one.
Commerce has quietly paid for the web for 25 years. AI is about to decide where that money goes next.
Every search ad, sponsored post and affiliate link is a seller paying to reach a buyer, and that money kept the web free. Now AI answers the question before anyone clicks. The money isn’t going away; it’s moving into the answers, and less of it reaches the people who made the web worth searching. An open commerce layer lets it flow back: openly, fairly, and to everyone who helps people decide.
One shared layer between buyers and sellers.
Everything starts with an intent: what someone needs, stated by them or by an agent working for them. The product graph finds what fits, sellers respond with offers, and the protocol matches them on merit. Inomy is the buyer app and Merx is the seller portal. Both run on this layer today.
Three components make it work.
Each one is shared infrastructure: built once, open to everyone.
Every listing, resolved into one clean record any agent can read.
Answers: product knowledge is scatteredExplore the graph ↓02Proves the rules heldTrust layerRules written into the protocol, and proof for anyone that the market played fair.
Answers: every recommendation is paid for somehowSee how trust works ↓03Shares value openlyOpen economicsValue shared in the open, and never allowed near the match.
Answers: the value stays in a few placesSee the economics ↓Every product, one open record.
The same laptop is described a dozen ways across a dozen stores, and every platform keeps its catalog to itself. Agents can’t reason over that mess. The open product graph resolves every listing into one clean, structured record: standardized, deduplicated and enriched, so any agent can read and compare it.
It’s live and growing, and soon any seller will be able to list in it and any agent to read from it.
Fair by design. Provable by anyone.
Ownership, matching and settlement run on rules written into the protocol, not on anyone’s policy. And the protocol keeps the receipts: who owns what, that matching stayed neutral, that every fee settled as it should.
Anyone can check the proof. No one sees anyone’s purchases.
The first pieces run today, and each phase makes more of the record public.
The goal.
Every market run will read like this, and anyone will be able to check.
Economics built in, so it stays open.
Protocols without an economic model end up owned by whoever monetizes them. So the economics are part of the protocol, bound by three rules: every fee in the open, no money near the match, and value for everyone who helped.
The mechanics will be set in public, with the market.
The first applications.
Two applications, one on each side of the market, prove the design and shape the protocol. We build them with Allen, our open-source agent OS. All three are live today.
Inomy. A shopping assistant built on the protocol’s product graph and matching: the buyer-side application. You can use it today.
Try Inomy ›LiveMerx. The seller portal, where retailers, marketplaces and brands bring their products to the protocol and meet demand from buyers and their agents.
Visit Merx ›Open sourceAllen. The open-source agent OS we build Inomy, Merx and the product graph with. Anyone can use it to build their own applications on the protocol.
Explore Allen ›What anyone will be able to build.
Anyone will be able to plug in through open APIs, and because the economics are part of the design, each of these can earn from day one. The APIs open in phases. Get on the list →
Shopping or general-purpose assistants, like Inomy, that earn from the commerce they help with.
Tools that read declared demand: what buyers actually ask for, and where supply falls short.
Publish reviews and guides, and get credited when they help someone decide, whichever seller they choose.
Specialist agents for any category deep enough to deserve one.
An intent becomes something you can program: hold it open, set rules for how it’s met, bid, or buy as a group.
Reviews tied to purchases the protocol can verify: feedback that’s far harder to fake.
ⓘ Scenes are illustrative: they show what the open APIs will make possible.
Fair questions.
Inomy and Merx are live today, with the first protocol pieces behind them. Open APIs, the economic model and the public record are what we’re building next.
Sellers, just as they fund marketing today, through fees defined in the open. The exact model will be set in public, with the market. Buyers never pay, and money never touches matching.
No. Money never decides what gets recommended: the product is chosen on fit before any money is involved, and every fee is defined in the open.
Selection happens before money enters: the product is chosen on fit alone, then sellers of that product compete on their offers. That separation is the core design rule, and it will be enforced in code.
Because curation always bends toward whoever pays for it, through ads, commissions or store brands. Open rules anyone can check keep money out of the match, the way open protocols kept email and the web open to everyone.
They’re real progress: they standardize how agents talk to stores, check out and pay. They don’t define what a product is across sellers, how offers are ranked, or how value is shared. That’s the layer we’re building, designed to work alongside them.
Proofs that the rules were followed: ownership, neutral matching, settlement. Purchase details stay private. The public record is one of the pieces we’re building now.
We do, while the first pieces are built. It’s designed so that no single company, including us, controls it, and it opens up in phases. If you want early access, talk to us.
This is how commerce should work. We’re building it.
The first pieces are live: try Inomy, meet Merx, explore Allen. If you’d like to build on the protocol, talk to us.