Canvas takes over the layout and composition half of a Shopify build. It does not take over the data modelling, the Liquid logic, the custom section schema or the integrations, and anyone telling you it does has not shipped a store on it yet. Shopify Canvas went live on 1 October 2026 as an infinite design workspace for the online store, and we have now run it on enough builds to say with some confidence which jobs genuinely leave the theme editor and which ones still land back in the code editor by Thursday.
The short version: if your theme is built on theme blocks, Canvas is the better surface for roughly everything a designer does after the wireframe stage. If your theme is a 2019-era vintage build with a 4,000-line product-template.liquid, Canvas will not save you. It will just show you, on one screen, how much of the store is hardcoded.
What Canvas is, and how it actually behaves
The theme editor was a single-template view. You picked a template from a dropdown, edited sections in a left rail, previewed in an iframe, saved, then picked the next template. Canvas drops that constraint. You get a zoomable workspace with multiple templates laid out side by side, and you edit blocks directly on the surface rather than through a form in a sidebar.
Mechanically it is still writing to the same place. Section and block settings land in the theme's JSON template and settings files, the same as before. Nothing new is being stored. What changed is the unit of work: instead of "open the product template, find the section, change the setting", it is "find the thing on screen, change it, see every other template that uses it update at the same time".
That last part is the whole product. Everything else is a nicer interface.
Shopify Canvas vs the theme editor: the difference that pays for itself
Here is the arithmetic from a build we were mid-way through when Canvas landed. Client came back after the first review wanting the primary button changed: different corner radius, different hover state, slightly tighter padding. That button appeared on six templates.
Old path: open each template, locate the sections using the button, change the setting, save, move on. Call it three minutes per template if nothing surprises you, so eighteen minutes, plus a ten-minute pass across the store to catch the two places you missed. Twenty-eight minutes for a corner radius. Then the client changes their mind once and you do it again.
In Canvas, because the button was a theme block with a shared preset, it was one edit. The six templates updated in the same view while we watched. Under two minutes including the client agreeing on the number.
The catch, and it is a real one: that only worked because the theme was built properly. On a different store the same month, the previous developer had styled buttons inline inside three separate section files. Canvas showed all three side by side and we could see they were inconsistent, which was genuinely useful, but fixing them still meant opening the code editor. Canvas made the problem visible. It did not make it editable.
Can designers edit every page at once?
Every page that shares a block, yes. Every page in the store, no, and you would not want that anyway.
Think of it in three tiers. Theme settings, like your type scale and colour schemes, have always been global and still are. Shared theme blocks propagate wherever they are used, which is the new behaviour and the reason Canvas feels fast. Template-specific overrides stay local, so the hero on your festive landing page does not inherit the hero on the homepage unless you built them from the same block.
Where designers get caught out is assuming a visual match means a structural one. Two sections that look identical but were added separately are two separate things, and editing one will not touch the other. We spend the first hour of any Canvas handover walking the designer through which elements on the board are actually the same object. On a theme someone else built, we get that mapping wrong on the first pass maybe a third of the time.
Does Canvas work with your current theme?
The gate is theme architecture, not your plan.
Themes built on the block-based architecture that Horizon introduced get the full experience: cross-template editing, drag-anywhere composition, nested blocks, the lot. Horizon and its variants were designed for this surface, and Canvas plus a Horizon-family theme is the combination Shopify clearly has in mind.
Older Online Store 2.0 themes, the section-and-JSON generation, open in Canvas and are workable. You can see templates side by side, you can rearrange sections, you can edit settings. What you lose is the shared-block propagation, because the theme was never built with shared blocks. You end up doing the same number of edits as before in a prettier window.
Vintage themes, the pre-2.0 ones still running on a surprising number of Indian stores that launched around 2018-2020, are a different conversation. If you are on one of those, Canvas is not the reason to move, but it is one more item on a list that already includes app-block support, metafield-driven content and the fact that nobody is writing documentation for your theme anymore. We cover what that move involves in theme and platform migration.
One honest note: a theme refactor to get block-based architecture is not a weekend job. On a mid-size catalogue store we budget it like a rebuild, because that is what it is. Do not let "we want Canvas" be the business case on its own.
Which plans get it
It is an online store feature, so it is not gated behind Plus. Basic upward, same editor. If you were hoping Canvas would be the thing that finally justified the Plus upgrade, it is not. The Plus case is still checkout extensibility, B2B, Functions, multiple stores and the API limits. Design tooling is not on that list.
Shopify Canvas limitations: where you still open the code editor
Canvas composes what exists. It does not create new capability. Everything below still landed on a developer in the last few months:
- New section or block types. If the behaviour you want is not in the theme's block library, somebody writes Liquid and a schema. Canvas then places it beautifully.
- Conditional logic. "Show the COD-available badge only on PIN codes we serve", "hide the EMI widget below ₹3,000", "swap the size chart by product type". That is Liquid, metafields and sometimes an app. None of it is a layout decision.
- Checkout. Separate surface, separate rules, Plus-dependent for the deeper customisation. Canvas stops at the storefront.
- Data modelling. Metaobjects, metafield definitions, taxonomy, how your ingredient list or BIS hallmark data is structured. Canvas renders it. Someone still has to design it.
- Anything involving a third-party script. App embeds behave the same as they did. Your Razorpay widget, your review app, your WhatsApp button, all unchanged.
That list is not shrinking fast. What Canvas removed was the tedium around the work, not the work.
The speed trap nobody is talking about yet
Making composition easy makes adding things easy. We have already watched one store go from a clean 2.1s LCP on mobile to 3.4s over three weeks of enthusiastic in-house editing, because each addition cost nothing in effort and the hero gained a video background, two badge rows and an above-fold testimonial carousel that loads a font of its own.
Nothing about Canvas caused that. The absence of friction did. If your team is going to be editing daily, put a number on the wall: a mobile LCP ceiling, a maximum page weight, whatever you will actually enforce, and check it weekly rather than at the end of the quarter. We use SwiftStore to keep an eye on the score between releases on stores where the client edits their own storefront, and we still do a proper performance pass before every festive peak. September is not the month to discover your product page gained 900KB in July.
Running a design review in Canvas with a client
This is where Canvas changed our process most, and it is worth being deliberate about it.
Work on a duplicate theme, never the live one. Obvious, but Canvas makes editing feel so lightweight that people forget which theme they are in. Name it with the date.
Populate it with real products before the client sees it. The whole value of an infinite board is seeing the system, and a system looks fine with three placeholder SKUs and falls apart with 180 real ones where half the titles run to two lines and one supplier sent 900x900 images. Load the real catalogue first.
Lay the board out by journey, not by template type. Home, collection, product, cart, in that order, left to right, with the mobile view of each directly beneath. Clients read it the way they read a storyboard and the conversation stops being about individual pages.
Give one person edit access. Not the founder, the brand lead and the agency designer simultaneously. Everyone else comments. We learned this the hard way on a build where two people repositioned the same block in opposite directions over a lunch break.
Set a window. Two working days to leave comments, then the board freezes and changes go into the next round. Canvas removes the friction that used to limit revision rounds naturally, so you have to put the limit back in on purpose, or the project has no end.
What it means for build budgets
Take a storefront build quoted at 120 hours. In our experience the layout-churn portion, the revisions after the first design review, is the part Canvas compresses, and it compresses hard, sometimes to a third of what it was. But that portion was never the bulk of the quote. Discovery, data modelling, custom sections, integrations, performance work and QA are untouched. So a 120-hour build might come in at 105. Useful, not transformative.
What it genuinely changes is the ongoing relationship. Merchants who were paying an agency retainer for small layout changes can do most of those themselves now, assuming the theme was built on blocks. That is a good outcome and we would rather say it plainly. Spend the retainer on things that compound instead: catalogue structure, search and filtering, speed, content and SEO. If you want the current numbers on what each slice of a build costs, we keep them here: Shopify development cost in India.
Where to start
Open your current theme in Canvas and try to change one button everywhere. If it propagates, your theme is in good shape and you should move your design reviews onto the board this month. If you end up in six different templates, you have a theme architecture question to answer before you have a Canvas question.
If you want a second opinion on which of those two you are in, send us the store URL and we will look at the theme structure as part of a free audit. Takes us about a day.