PlatformDTC Is in Public Beta. Build a Store Today, Publish It for $1.

As of today, PlatformDTC is in public beta. A stranger can sign up at platformdtc.com/register, and at the end of it there is a real store on a subdomain — an account, a storefront and a default theme already seeded — in the same dashboard the merchants already trading here use. No sales call, no approval queue, no invite code, no waitlist email that arrives four days later.

What sign-up does not do is put that store in front of shoppers. Building is open to anyone: the theme editor, the catalog, the checkout settings, all of it on the real product rather than a demo of it. Publishing is what a plan buys — the storefront starts serving the public when you start one, and the first month of a plan is $1.

That is the whole announcement. Everything else in this post is detail about what you get and what we are still willing to break.

What “public beta” means, stated plainly

The phrase is used loosely enough to be worthless, so here is the specific version we mean by it.

• You can sign up today. There is no gate in front of building a store. The gate is on publishing one: the storefront goes live when a plan starts, and each plan covers one store.
• Things will change under you. Screens move, defaults change, and behaviour we think is wrong gets fixed rather than preserved for compatibility. When it is the kind of change a merchant would notice, it goes in the public changelog with what changed and what kept working.
• Telling us something is broken reaches a person. support@platformdtc.com is read by a human, and the full department list is on our contact page.
• You can check us instead of trusting us. status.platformdtc.com publishes the live state of the surfaces we probe — storefronts, checkout, dashboard, API, notifications, docs — with p50 and p95 response times per component, not just a green dot. Two components, background jobs and analytics, are not probed yet and say so on the page rather than showing a green dot they have not earned.

What beta does not mean here: it is not a demo environment, the stores are not sandboxes, and the orders are real orders. Merchants are taking live payments on this platform now.

What is actually covered

The reason to open sign-up at all is that the surface is wide enough to run a store end to end rather than as one piece of a stack you assemble yourself. In one account:

• Storefront and themes, product catalog, and the editor that changes them.
• Checkout, payments, and subscriptions.
• Fulfillment and shipping, with branded order tracking on your own store.
• Email and SMS marketing.
• Analytics.

This is a list of things that exist and are being used, not a roadmap. Where a piece is thin we would rather you find out from the product than from a bullet.

Checkout runs on your domain, not ours

Attach a custom domain and checkout runs on it. The shopper who clicked Buy on yourbrand.com pays on yourbrand.com and never sees our name in the address bar. Before you attach one, checkout runs on the platform domain at platformdtc.com/checkout/<session> — the same origin as everything else, so the highest-intent step of the funnel does not pay for a fresh DNS lookup and TLS handshake.

Express wallets available at checkout: Apple Pay, Google Pay, Link, Amazon Pay and Klarna. Wallets only render on a hostname registered with the wallet provider, which is why this is a short factual list rather than a logo wall.

What it costs

Three plans, each priced per store. The price is the store, which is why there is no store count in the table — a second store is a second plan.

The first month is $1 on whichever plan you start with, once per account. Card processing is charged per transaction at the rate for your plan. Every new merchant also gets a one-time grant of $100,000 in AI credits, which exists because the agent surface below is the part we actually want you to use, and metering it at zero for a while is cheaper for us than watching people not try it.

Building before that first dollar is not a trial and there is no clock on it — but it is also not a published store. Nobody can buy from a storefront that has not gone live, and going live is the thing the plan pays for. We would rather say that in a sentence than let you find it out on the day you send your first campaign.

The part that is different

A store builder with a chat box in the corner is not what this is. Capabilities here ship as skills, and a skill takes real actions in real systems — it publishes the actual theme, calls the actual ad platform, sends the actual campaign. There is no simulation layer.

That is only a good idea if the agent is contained by construction, so agents reach the platform through the Agent Gateway: a scoped key per agent, an audit record for every call, an idempotency key on every write so a retried request cannot double-charge, and an approval gate that parks spend until a person resolves it.

POST /campaigns/{id}/send          # approval-gated key
→ 202 { "status": "pending_approval", "job_id": "…" }
# nothing has been sent. a human decides.

All of it is documented publicly, including an OpenAPI reference for the Gateway, at docs.platformdtc.com. You do not need an account to read it.

What we are not claiming

Earlier this month we published two measurements of other people’s stores: 70 of 147 DTC brands serve a live agent-commerce discovery document, and 31 of 58 well-known storefronts answer a Model Context Protocol request with 13 schema-d commerce tools. It would be convenient to launch a public beta and imply we are in that set. We are not. PlatformDTC storefronts do not currently serve UCP or agentic-checkout endpoints — request them and you get a 404. We measured the gap on other people’s stores; the same method applied to ours returns the answer we do not like.

Saying so is not modesty. Anyone can run the check in about ten seconds, and a launch post that skipped it would be a claim with a public disproof sitting one curl away.

Self-serve is a second door, not a replacement

Earlier this month we moved the public surface upmarket on purpose: the primary call to action is Talk to sales, and it stays that way. If you are migrating a real catalog, moving live subscriptions, or need someone to look at your actual numbers before you commit, a conversation is a better use of your time than a sign-up form, and it is the door we would point you at.

Public beta adds the other door. Some people would rather see the product than be told about it, and until today there was no way for them to do that. Now there is.