It's been six months since UCP was announced, and the ecosystem census now reads like a protocol winning its first argument at speed. The fleet keeps growing — fast, and for the fifth straight month. The payments industry spent June and July shipping agent-payment rails at a furious pace — the clearest signal yet that the biggest names in the industry are converging on this future. What remains is the connection between the two: the stores those rails are built to serve haven't switched payment on yet. That handoff — the moment the fleet plugs into the rails — is now the single most important event coming in agentic commerce, and this month's census shows just how much is already staged for it.

This is the sixth monthly UCP ecosystem report. Every number below is observed from the outside — what stores actually publish and how they actually respond — not what anyone claims. Here's where the ecosystem sits.

#The numbers

UCP crossed 11,414 verified stores this month, up from 8,259 in June — a +38% jump, and the fifth straight month of compounding growth. Total domains monitored reached 13,037, with 1,920 new merchants in the last 7 days and 4,470 in the last 30.

Figures are a point-in-time census as of July 7, 2026.

The version story has effectively resolved: 11,384 of the 11,414 verified stores — 99.7% — run spec version 2026-04-08. The long tail of 2026-01-* drafts has all but disappeared. That's a remarkable outcome for a six-month-old protocol: the ecosystem is standardized on a single spec revision, which makes everything downstream — conformance, capability comparison, transactability — an apples-to-apples question. Very few protocols achieve this kind of version discipline this early.

Growth is not the interesting story anymore. It's steady, it's broad, and it's going to keep going. The interesting story is what all these new stores can — and will soon be able to — do.

#The capability cliff

Here is the capability histogram for the fleet, and it is not a gentle slope. It is a cliff.

• Checkout — 11,389
• Order — 10,114
• Cart — 10,109
• Identity — 15
• Payment — 0

Read that last line again. Zero. Six months into UCP, across all 11,414 verified stores, not one yet exposes a payment capability an agent could complete a purchase against. Checkout is declared nearly universally; payment — the step that turns a declared ecosystem into a transacting one — hasn't switched on yet.

This is the single most important number in agentic commerce right now — and, read correctly, one of the most encouraging. Eleven thousand stores have already done the integration work: discovery, cart, and checkout are live, conformant, and standardized on one spec revision. An agent visiting the median UCP store can discover products, build a cart, and walk right up to the checkout. The fleet is staged at the payment line — waiting, sensibly, for the rails and the trust layer to settle. When that last field flips, it flips against an installed base that already exists. Identity, the capability that lets an agent prove who it's acting for, sits at 15 stores — and a further three declare identity-flavoured capabilities in vendor or non-standard namespaces, early stirrings on the same seam. The zero isn't a void; it's a coiled spring.

Two footnotes make that zero more encouraging than it looks. First, payment through UCP has already been proven end-to-end at runtime: in March, the first fully autonomous agent purchase completed a real order — wallet payment included — at a live WooCommerce store. The plumbing works; what the census counts is who declares payment in a manifest as of the cutoff. Second, the first populated payment_handlers blocks are starting to appear at the edges of the ecosystem: a live travel-vertical implementation at ucp.travel publishes a provider-payment handler with card and balance instruments — and declares ap2_mandate — in its manifest today. It sits outside the retail census above, and its handler lives in a vertical extension namespace (exactly the kind of declaration the authority-bound map below exists to verify), but the direction is unmistakable: the payment line is getting ready to move.

We've called this the reliability gap in past reports — the distance between what a store declares and what it can actually do. This month it stopped being a gap and became a cliff, and the cliff is at the payment line. You can watch it yourself, live, at ucpchecker.com/capabilities.

Scoring tells the same story from another angle. Of the 1,183 stores scored so far, the average UCP Score is 56.3 out of 100 — a solid C. The distribution shows why: a big middle of B and C grades (stores that get discovery and cart right), 9 A-grade stores, and a tail of 289 failures — stores that declare UCP but don't yet hold up when you actually knock. The 9 A-grades matter more than their count suggests: they prove full conformance against today's spec is achievable right now. The ceiling is real, and it's reachable.

#Authority-bound: proving an extension belongs to who it claims

The other thing that happens as an ecosystem matures is that it sprouts extensions — vendor-specific capabilities layered on top of the core spec. That UCP is already sprouting them is itself a healthy sign. And the moment that happens, a new question appears: when a store declares an extension, how do you know that extension actually belongs to the vendor it names? Anyone can write com.somevendor.loyalty into a manifest. Proving the vendor stands behind it is a different matter.

This month we shipped the answer: an authority-bound vendor & extension map, live at ucpchecker.com/vendors. It maps each extension back to the authority that defines it and verifies the binding — so a declared extension is either proven to belong to its vendor or flagged as unresolved. We're tracking 5 authority-bound capabilities across 5 extensions at launch, and the map grows as the ecosystem does.

Why it matters: as agents start trusting extension-declared capabilities to make decisions — which payment vendor, which loyalty program, which fulfillment authority — an unbound extension is an open door. Authority binding is the difference between "this store says it supports Vendor X" and "Vendor X's authority confirms it." It's the same discipline as message signatures, moved up a layer: don't trust the claim, verify the binding. It's the piece of infrastructure the ecosystem will need the day extensions start carrying real weight — which, given the direction of travel below, is coming.

#The rails are ready and waiting

Now set that capability histogram against what the rest of the industry did in June and July, because the contrast is the whole story — and it's a story of conviction.

• Mastercard launched Agent Pay for Machines — infrastructure purpose-built for high-frequency, low-latency, agent- and machine-executed payments.
• Open USD arrived on June 30, with Stripe, Visa, Mastercard, Coinbase, BlackRock, Google, Shopify and 140+ others behind a dollar-backed stablecoin — a settlement rail for exactly this kind of transaction.
• Stripe and OpenAI's Agentic Commerce Protocol continued powering Instant Checkout inside ChatGPT.
• PayPal and Perplexity's Instant Buy is live in production, making PayPal merchants purchasable from inside an AI answer engine.
• Visa's Trusted Agent Protocol is public on GitHub and the Visa Developer Center, with AI partners including Anthropic, OpenAI, Mistral and Perplexity.

And in the spec itself, the live work is precisely on the payment-and-identity seam: the ap2_mandate payment-authorization flow (which requires ES512, a signature algorithm the current algorithm table doesn't yet define), the exact byte-encoding of signatures (raw r‖s versus DER), a full 3D Secure flow under active deep-dive, and a Technical Council decision to carry saved payment methods on identity-linking infrastructure rather than a new extension — which makes the identity-linking count above the quiet leading indicator of transactability. The people writing the protocol are hardening the payment handoff right now — exactly the work that has to land before the fleet can flip payment on responsibly.

So the picture is unambiguous, and for anyone invested in agentic commerce it should read as momentum, not warning. Every major payment network, wallet, and agent platform is racing to make the moment of purchase work. The fleet isn't wired into those rails yet — but that's sequencing, not absence of demand, and the demand side has never looked stronger. The rails converging on UCP-shaped commerce is the strongest external validation the protocol has received to date.

That the stores haven't connected yet is the natural order of a rollout, not a flaw in the protocol. Discovery and cart are the easy 80%; wiring live payment is the hard, high-stakes 20% that carries real money and real liability, and stores are — reasonably — waiting for the rails, the standards, and the trust layer to settle before they turn it on. The headline adoption numbers describe a declared ecosystem on its way to becoming a transacting one. The day the payment line moves off zero is the day agentic commerce goes live at scale — and everything in this report says that day is getting closer.

#The platform wave starts: Wix shows up

For five months, the single highest-impact event in UCP adoption was the one that kept not happening: a website platform turning UCP on across its whole fleet. Individual merchants adopt one at a time. A platform flips tens of thousands at once. SFCC, Adobe Commerce, Squarespace, Wix — all quiet.

This month, one of them moved. Wix has started shipping UCP. We're seeing Wix-hosted manifests served through a centralized gateway — www.wixapis.com/ecom/ucp/{store-id}, exposing both REST and MCP transports — which is the fingerprint of fleet-level infrastructure, not a hand-rolled one-off. Wix's own Base44 stores are already turning up in the crawl carrying versioned 2026-04-08 manifests with a fuller capability set: checkout, fulfillment, and catalog lookup and search.

This matters more than any single merchant signing on, because of how platform rollouts move. Once the gateway is wired, the manifest is a template every store on the platform can inherit at once — so a platform-driven curve doesn't add hundreds a week, it adds tens of thousands in a single step change. Wix appearing is the first real sign that the platform wave everyone's been waiting on has actually begun to break.

With one very consistent caveat: those Wix manifests declare shopping and carry an empty payment_handlers — the same staging pattern as everyone else. The template is live; payment is the field left to flip. The day a platform of Wix's size flips that field across its fleet, it won't be a data point in next month's report. It'll be the month the whole picture changes.

#New this month: from a check to a watch

Everything in this report is a snapshot — and if six months of censuses prove anything, it's that the picture moves. Manifests break on deploys, capabilities appear and disappear, and a store that was agent-ready on Monday quietly isn't on Thursday. That's why this month we opened early access to the UCP Checker monitoring dashboard: daily re-checks of your own store against the live spec, break alerts the moment agent access degrades, UCP Score trends over time, and benchmarking against the stores on your watchlist. Verification is one DNS record, and it's free during the beta.

Further reading: Introducing UCP Monitoring: Status, Uptime, Score & Alerts in One Dashboard

#What to watch in August

• The payment line moving off zero. This is the one that matters. The first stores to publish a real payment capability — most likely platform-driven, arriving in cohorts rather than one at a time — will mark the moment the cliff gets a floor. We'll be watching it daily.
• Wix flipping payment. Wix's fleet template is live and declares empty payment_handlers; the month it turns payment on, transactable UCP grows in tens of thousands, not ones. It's the highest-leverage number on the board.
• AP2 mandates landing. If the payment-authorization flow and its signature requirements settle in the spec, the technical blocker to that first payment cohort comes down with it.
• Payment terms enter the spec. A new RFC (#587) adds deposit-now/balance-later, net terms, and buy-now-pay-later to checkout — the first structured work on how money moves, not just whether. Watch whether it ships before the first payment cohort does.
• The authority-bound map filling in. As extensions multiply, watch how many resolve to a verified authority versus how many stay unbound — it's an early read on how much of the extension layer is real.
• Discovery gets a working group. Meta opened an RFC (#550) for a bulk product-discovery layer — pre-generated feeds that let platforms ingest a merchant’s full catalog without crawling — and the Technical Council has agreed to charter a domain working group to settle feed formats, diagnostics, and row-level data control. It’s the first real motion on the discovery end of the funnel, the layer that decides whether the long tail gets found at all. The thing to watch: whether it lands as an open, public tier or an authenticated, platform-gated one — early signals point public, with the RFC scoping feeds to public, non-sensitive product data.
• The August release train. The Technical Council minutes — now public through late June — show a quarterly release cadence agreed, with the next release targeted for August 2026 and v0.8 planning centered on 3D Secure, Web Bot Auth interop, and fulfillment, aimed to land before the holiday shopping season. The protocol now has a shipping schedule.
• UCP goes multi-vertical. Domain Technical Councils are forming: Food Ordering nominations are open, the Lodging charter is live, a Payments charter is expected, and a repository refactor moves payment constructs from the shopping vertical into common. The travel-vertical manifest in this report's payment footnote is early field evidence of the same direction — we track each vertical's fleet at ucpchecker.com/verticals.

Six months in, UCP has won the argument that had to be won first: stores are adopting it, fast, and standardizing as they do. The next argument — an agent completing a purchase at the median store — is the one every force in this report is now pushing toward: the rails are live, the spec work is on exactly the right seam, and the first platform template is already staged at the line. Zero is a number with an expiry date. We'll see you in August.

#About UCP Checker

UCP Checker is the independent validation and observability layer for the Universal Commerce Protocol. We crawl, validate and grade every public UCP manifest we can find, run the merchant directory, the UCP Score and live adoption stats, and track the spec as it evolves so you don't have to. For businesses, the UCP Checker app turns the one-time check into continuous observability — daily conformance monitoring against the live spec, break alerts the moment agent access degrades, and benchmarking against the stores you compete with. The capability cliff in this report is exactly what a neutral, outside-in meter exists to surface — what stores actually expose, measured the same way for everyone, without picking winners.

• Check your store: ucpchecker.com/check
• Grade it (UCP Score): ucpchecker.com/score
• Monitor it continuously (UCP Checker app): app.ucpchecker.com
• Capability coverage: ucpchecker.com/capabilities
• Authority-bound vendor & extension map: ucpchecker.com/vendors

#Sources

All data is from the UCP Checker crawler, which re-checks every tracked domain at least every 24 hours, as of 7 July 2026. The verified-merchant dataset is published monthly on Hugging Face under CC-BY 4.0; the same data, a public REST API, the bulk checker, and the rest of our developer tools are ungated.

• Capability coverage: ucpchecker.com/capabilities
• Authority-bound vendor & extension map: ucpchecker.com/vendors
• Browse the directory: ucpchecker.com/directory
• Track adoption live: ucpchecker.com/stats
• Run a UCP Score: ucpchecker.com/score
• UCP Technical Council minutes: meeting-minutes/tc/2026
• Previous report: State of Agentic Commerce — June 2026

External coverage cited in this report:

• Mastercard launches Agent Pay for Machines, Mastercard, June 2026
• Stripe powers Instant Checkout in ChatGPT and releases the Agentic Commerce Protocol, Stripe
• New tech and tools for retailers to succeed in an agentic shopping era, Google (Ads & Commerce)
• PayPal and Perplexity launch Instant Buy, PayPal Newsroom
• Open USD: Stripe, Visa and 140+ businesses launch a shared stablecoin, Fortune, June 30 2026
• Visa Trusted Agent Protocol: GitHub · Visa Developer Center
• UCP spec work cited: #571 ES512 algorithm-table gap · #569 signature byte-encoding · #587 payment terms RFC · #550 bulk product-discovery RFC