A multi-tenant platform for print, from design to delivery

Each client lives in its own account with its own users, template access, and quotas. Jobs and stored designs carry the account ID, and every list view — orders, projects, usage — is filtered to the signed-in account. There is no shared bucket of data to leak across.

Clients log in with a username and password and get a signed ps_session cookie. The middleware verifies it at the edge; API routes re-check it against the auth store on every call. First login forces a password change from the admin-issued temporary one.

The env-var super-admin runs everything from /dashboard/accounts: create accounts, set permissions and quotas, curate template access, add or suspend users, and watch per-account usage — all from one place.

Every template is built at 300 DPI on a page-sized canvas (A4, Letter, or any custom size you configure). Bleed and safe-zone guides are drawn on the canvas, so designers don’t need to remember to leave room — the boundaries are visible as they work.

The viewport zooms and pans like you’d expect. Undo and redo are unlimited inside a session.

The Layers panel shows everything on the canvas as a tree. Layers can be renamed, locked, duplicated, hidden, or grouped. Right-click any layer for the full set of actions.

Groups have a dedicated edit mode — double-click in to tweak a child without losing the group’s overall position. Clip masks are first-class: any group can be turned into a mask so a photo can be cropped to a circle, a logo shape, or any custom path.

Full text editing on canvas. Pick from a curated set of typefaces (Inter, Outfit, Anton, Playfair, Roboto, and the rest of the loaded library) and adjust weight, size, line height, and tracking inline.

The color picker is CMYK-aware — the value you set is the value that ends up on the printed page. You can also bind any text’s color to a palette role (primary, secondary, tertiary) so it picks up whatever color is chosen at order time.

Rectangles, circles, lines, polygons, and arbitrary paths. Each shape has an independent fill and stroke, both with the same CMYK-aware color picker as text. Shapes can also be bound to palette roles for order-time recoloring.

When you drag a layer, the canvas snaps to the visual edges of nearby layers. Text snaps by glyph bounds — meaning two stacked names line up by their visible letters, not by their invisible textbox borders. Same with logos that have whitespace baked into the file.

Select any layer in the studio and give it a tag — usually something descriptive like name, phone, headshot, or logo. The tag is the handle the order form uses.

Text, image, and shape layers can all be tagged. The tag panel shows everything that’s tagged in the current template at a glance.

Shapes and text can be assigned a palette role: primary, secondary, or tertiary. When the template is ordered, the chosen brand palette flows into every layer that uses that role.

One palette change per order, not a per-layer recolor. And because roles are assigned at design time, a designer can guarantee the brand colors land in the right places.

Each image field carries its own fit mode and anchor. So if “headshot” is set to cover, top-center, every photo swapped in at order time will crop the same way — heads stay above the fold, no awkward chins or foreheads.

As you tag layers, the studio maintains a manifest of every dynamic field — the tag, the type (text / image / color / shape), the constraints (max length, image fit, palette role). This manifest is what the order form reads to decide which inputs to show.

Designers don’t build forms. The form is the manifest, rendered.

Every order page shows live status as the job moves through the pipeline: queued, template loaded, layout built, PDF rendered, color converted, done. The percentage ticks up as each stage completes.

The dashboard gives the team a system-wide view: how many jobs are active right now, what’s queued, which jobs failed and why, and how the template cache is performing.

It’s where someone running operations would start their day to see if anything needs attention.

Every job is logged with its inputs and outputs, including failures. If something fails — bad image URL, transient network error, missing asset — operators can fix the underlying issue and retry the job from its history entry without re-entering the form.

Final PDFs are uploaded to an AWS S3 bucket organized by job ID and date. Each finished output gets a stable public URL that can be shared, downloaded, or piped into a downstream system without going back through Print Studio.

Alongside each PDF, the system also saves an editable .pst of exactly what was rendered. That means any finished card can be reopened in the studio later — to tweak a color, fix a typo, or use it as the basis for a new design — without starting over.

A small WebP preview is saved with each output, sized for fast scanning in lists. Useful for picking a recent order out of a history view at a glance, without needing to open the full PDF.