The 60-second pipeline: from logo to live merch storefront

For most brands, the gap between wanting a merch program and having one is six months and a few thousand dollars of unsold hoodies. We built Merch365 because that gap is mostly artificial — almost every step in it is automatable, and the parts that aren’t are already solved by an existing decorator network. Here’s what actually happens in the sixty seconds between the logo upload and the live store.

Step 1 — Logo normalization (~5 seconds)

You drop an SVG, PNG, or JPG. The pipeline doesn’t trust any of them.

Vectors get re-flowed. SVGs from Illustrator carry pasteboard junk, hidden layers, and CMYK color profiles that break on the press. We strip down to a minimal path set, convert to RGB, lock the safe-area, and emit a clean working copy.
Rasters get reconstructed. PNGs and JPGs go through edge detection + background subtraction, then we generate a vector trace at print resolution. The original raster stays as a fallback.
Variants get rendered. Light, dark, mono, knock-out, square crop, badge, social — seven canonical forms. These aren’t decorative; they’re the inputs the rest of the pipeline assumes are present.

The output is a brand kit — a content-addressable bundle that every downstream stage references by ID.

Step 2 — Catalog selection (~10 seconds)

We carry the full SanMar / S&S / Alphabroder catalog: about 8,500 blanks. No store ever shows 8,500 blanks. The model reads six signals from your brand kit — logo style, voice, audience, price tier, use case, seasonality — and ranks the catalog down to roughly 100 SKUs. The ranking is deterministic for a given input, which matters: when a prospect comes back to claim a store, the assortment hasn’t drifted.

The selection isn’t a recommendation. It’s the store’s actual inventory.

Step 3 — Mockup rendering (~30 seconds)

For every selected blank, we render a mockup with your mark in the right place. “Right place” sounds simple but isn’t: a chest hit on a tee is not the same coordinate as a chest hit on a hoodie, a beanie has one viable decoration zone, a cap has three. We carry a placement library per garment type, score each placement for readability against the blank’s actual color, and pick the winner.

The renders go straight into the storefront’s product grid. No theme. No image cropping in the CMS. The mockup is the product card.

Step 4 — Storefront stand-up (~15 seconds)

A subdomain reserves (yourbrand.merch365.shop), an SSL cert provisions, a Medusa sales channel scopes the catalog to your tenant, Stripe payment routing flips on, and the storefront image picks up the new tenant config from the edge. No deploy. No theme install. No app marketplace.

The fulfillment routing is already wired — when an order lands, it flows to the partner shop best positioned to print it. You don’t pick a printer; you don’t even see one. The packing slip ships under your brand.

What’s not in the sixty seconds

Two things people expect to be there, and aren’t:

Inventory commitment. Every order prints on demand. You don’t buy 200 tees on faith. If nobody buys, nothing prints.
A SaaS subscription. The storefront is free forever. We’re a decorator network — we make our margin on the printing, not the platform. If you never sell anything, we never bill you.

The whole model only works because we already do $5M/year in branded apparel through 50+ partner shops. The website is a way to send those shops more work. So it’s free.

What you do at the sixty-second mark

Approve, edit, or claim. That’s it. Most prospects we send a generated store to don’t change anything before going live — the assortment, the renders, and the brand application are right out of the box. The ones who do change things tend to swap one or two SKUs and adjust two or three retail prices. Then they share a link.