Medusa vs Shopify for multi-vendor marketplaces: what each platform makes you build
For a single-merchant store Shopify is the better answer, and it is worth saying so first. Marketplaces are where the comparison changes: Shopify Markets is one merchant selling to many regions, and Medusa's docs list order-splitting as something you write yourself.
For a single business selling its own products, Shopify is the better answer. It carries hosting, payments, PCI DSS compliance, a large catalogue of apps and an admin most teams already know, and taking that on yourself needs a reason beyond preference.
Multi-vendor marketplaces are where the comparison changes shape, and it is a narrower question than “which platform is better”. Several independent businesses each hold their own catalogue, prices and fulfilment. The buyer should see one storefront and pay once. The platform has to keep the sellers’ orders and their money apart underneath that.
This piece is about that case. WAYF built Kaffy, a multi-vendor coffee marketplace on Medusa, so the comparison below is written from having shipped one.
What multi-vendor requires
Four things have to be true before a site works as a marketplace.
One cart across sellers. A buyer adds items from three different businesses and pays once. Kaffy’s case study puts the distinction plainly: the single-cart pattern is the difference between a marketplace and a directory.
Order splitting after payment. That single payment becomes several orders, one per seller, so each seller sees and fulfils only their own items.
Per-seller fulfilment. Shipping is arranged and labelled per seller, because the goods leave from different places.
A seller-facing surface with its own permission model. Sellers need to manage their catalogue and their orders without seeing anyone else’s.
Every platform can be made to do this. What differs is where the seller model lives and who maintains it.
Shopify’s model is single-merchant-first
The feature closest to multi-vendor selling is Shopify Markets, and it solves a different problem. Shopify describes it as selling to multiple markets from a single store, covering localized languages, currencies, regional pricing, retail locations and B2B alongside D2C. The Markets page describes no independent-seller model; its one mention of marketplaces refers to selling through external channels.
Multi-vendor behaviour on Shopify is therefore added by third-party apps from its App Store, sitting on top of the merchant’s own store. For many marketplaces that is the right route, and it gets a product in front of buyers quickly.
Two published constraints are worth pricing before committing to it.
Checkout customization is tiered. Shopify’s pricing page lists checkout customization as limited on the Basic, Grow and Advanced plans, and a fully customizable checkout as a Plus feature. Plus starts at €2,100 per month. Order splitting is decided at and immediately after checkout, so a marketplace’s core mechanic sits in the part of the stack with the tightest constraints below the top plan.
Payment routing carries a fee when it is not Shopify’s. The same page states that third-party transaction fees may apply when a merchant uses a third-party payment provider instead of Shopify Payments, and that the rate depends on the plan: 2% on Basic, 1% on Grow, 0.6% on Advanced and 0.2% on Plus. Marketplaces frequently need a provider that supports split payouts to sellers, so this is a line item to check against the payout model rather than a detail.
Neither of these makes Shopify wrong for a marketplace; they define what an app-based marketplace can and cannot change.
What Medusa gives you, and what you write yourself
Medusa describes itself as an open-source commerce platform built to be customized, with the stated aim that you extend, customize and own every commerce workflow. The repository is MIT-licensed, apart from materials it identifies as Enterprise Edition. It can be self-hosted, and Medusa also offers a managed cloud, so self-hosting is a choice rather than the only path.
For marketplaces specifically, Medusa’s own documentation is unusually direct about the division of labour, and it is worth reading before choosing the platform rather than after.
Medusa provides the modular architecture. You create a marketplace module holding custom data models such as vendors or sellers, then link those models to existing ones like products and orders. It supports authentication for custom actor types, and workflows for multi-step processes with rollback and retry built in.
You write the rest. The documentation lists it: the custom data models, the API routes for vendor product and order management, the authorization logic that restricts those routes to authenticated vendors, a vendor dashboard or admin customization, the storefront, and the order-splitting itself. For multi-vendor purchases the instruction is to replicate the Complete Cart API route and create a workflow that splits the order into multiple orders, one for each vendor.
The seller model becomes part of your data model and your codebase, which is what makes marketplace-specific behaviour available to change. It also becomes yours to maintain, and an MIT licence does not reduce that. What the extra work buys is commerce logic that behaves the way your marketplace needs it to, instead of the way a platform and an app have settled on. That is worth paying for on a marketplace whose mechanics are its product, and it is overhead on one whose mechanics are ordinary.
What it looked like on Kaffy
Kaffy is a specialty coffee marketplace for Poland. WAYF designed and built it end to end: iOS, Android, web, the roastery panel and the admin, on a Medusa commerce engine, with a Payload instance alongside it and the database on PlanetScale for Postgres. Mobile is one React Native codebase compiled to both platforms, and the Next.js web build sits on the same API surface, so feature parity is a UI question. It launched on three dedicated servers and holds a 5.0 rating on the App Store in Poland.
The catalogue is a multi-vendor experience that presents itself as a single store. A buyer browses roasteries side by side, adds bags from three different brands to one cart, and pays once. Behind that, Kaffy splits the order across roasteries, generates the shipping labels for each, and returns a single order history to the buyer. The roastery panel’s permission model was specified in discovery, before any code, alongside the buyer’s first-run experience and the way multi-vendor checkout had to behave.
Medusa’s ecosystem also lists Mercur, an open-source marketplace starter that ships vendor panels and order splitting prebuilt — a real head start, though the code it gives you still becomes your team’s to evaluate, extend and maintain. Kaffy is what the order-splitting workflow looks like in production, with a client team running it afterwards.
Choosing between them
The choice turns on one question: whether your marketplace mechanics need to live in your own data model.
Shopify fits when one business sells its own goods; when a marketplace’s seller model can live inside an app’s data model and you are content for it to stay there; when time to market matters more than control of order logic; or when the team should not be maintaining commerce code at all.
Medusa fits when independent sellers each need their own catalogue, permissions and payouts; when checkout or order behaviour has to work in a way no platform offers; when the commerce data model has to be owned outright, which is also the vendor lock-in question; or when commerce is one part of a wider composable architecture rather than the whole product.
A marketplace on Medusa is a build, and the documentation says as much. If your team has, or can hire, the capacity to own that code, it buys behaviour you cannot otherwise specify. If it does not, Shopify with a marketplace app will be live sooner, and for most teams that settles it.
We work on Medusa builds through Medusa.js development. A free consultation and a scoped, fixed quote come before any commitment.
FAQ
Can Shopify do a multi-vendor marketplace? Not as a native capability. Shopify Markets, the feature closest to the idea, is about selling to multiple regions and channels from one store, and its page does not mention independent sellers anywhere. Multi-vendor behaviour on Shopify is added by third-party apps from its App Store, sitting on top of the merchant’s own store. That works, and for many marketplaces it is the sensible route. What you are accepting is that the seller model lives in an app’s data model rather than in yours, and that checkout, which is where order-splitting has to be decided, is only fully customizable on the Plus plan.
Is Medusa production ready? Yes, with one qualifier: Medusa is a framework, so what it ships is the commerce primitives, a modular architecture for adding your own data models, and workflows with rollback and retry for multi-step processes. The marketplace layer on top of that is code you write and maintain. WAYF built Kaffy on Medusa, a multi-vendor coffee marketplace running iOS, Android, web and a roastery panel on one API, and Kaffy’s own team runs the platform today. Production readiness is a question about your team’s capacity to own commerce code as much as it is about the framework.
What is the difference between Shopify Markets and a multi-vendor marketplace? Shopify Markets solves regional selling for one merchant. Shopify describes it as selling to multiple markets from a single store, with localized languages, currencies, pricing and B2B or retail variations, and its page does not mention independent sellers at all. A multi-vendor marketplace is a different shape: several separate businesses each hold their own catalogue, prices and fulfilment, and the platform has to keep their money and their orders apart while presenting one storefront to the buyer.
How does a marketplace split one order across several sellers? The buyer pays once, and the platform then divides that purchase into per-seller orders so each seller sees, fulfils and ships only their own items. In Medusa this is explicit work: the documentation directs you to replicate the Complete Cart API route and create a workflow that splits the order into multiple orders, one for each vendor. On Kaffy the same pattern produces one cart across several roasteries, a separate shipping label generated per roastery, and a single order history for the buyer.
When is Shopify the better choice over Medusa? Whenever one business is selling its own products, which is most stores. Shopify gives you hosting, payments, PCI DSS compliance handled by the platform, a large catalogue of apps and an admin your team already knows how to use, and none of that is work you should take on without a reason. The reasons to take it on are specific: marketplace mechanics with independent sellers, checkout or order logic that has to behave in a way the platform does not offer, or a requirement to own the commerce data model outright.
Sources
- Shopify pricing — the plan tiers, checkout customization listed as limited below Plus and fully customizable on Plus, Plus starting at €2,100 per month, and the third-party transaction fee rates of 2%, 1%, 0.6% and 0.2% by plan
- Shopify Markets — selling to multiple markets from a single store, and the absence of any multi-vendor or independent-seller claim
- Medusa, Marketplace recipe — the marketplace module, module links, custom actor-type authentication, workflows with rollback and retry, and the list of what the developer builds, including the order-splitting workflow
- Medusa — the open-source positioning and the stated aim of extending, customizing and owning every commerce workflow
- Medusa LICENSE — MIT, apart from the materials identified as Enterprise Edition
- WAYF, Kaffy — the stack, the single-cart and order-splitting behaviour, the roastery panel, and the App Store rating
Last verified against shopify.com and medusajs.com on 31 August 2026. Plan prices other than Plus are shown by Shopify in local currency and are omitted here for that reason; the transaction-fee percentages and the checkout-customization tiers do not vary by region.
We're booking content platform
engagements for 2026.
Twenty-five minutes to walk through the work and decide if we're the right team for it. Scoping and a fixed price come after.