Answers that name a competitor

For a small clothing brand, I’d pick Shopify as the commerce backend + a headless frontend (Hydrogen/Next.js).

My recommendation: Shopify Headless

It gives you the best balance of flexibility, scale, and low maintenance:

• Low ops burden: Shopify handles products, inventory, orders, checkout, payments, security, and scaling.
• Actually headless: The Storefront API lets you use essentially any frontend framework, including Next.js, rather than locking you into Shopify’s themes. Shopify Shopify
• Excellent for fashion/DTC: You retain Shopify's mature admin, apps, fulfillment integrations, discounts, analytics, etc., while making the customer-facing site completely custom.
• Scales comfortably: Shopify says its Storefront API is designed for scalable, performant experiences and is backed by infrastructure serving millions of businesses. Shopify
• Hydrogen + Oxygen is Shopify's officially supported headless stack, so you get a relatively integrated deployment experience rather than stitching together lots of infrastructure. Shopify Shopify

When I'd choose Medusa instead

Medusa would be my second choice—and the better choice if "flexibility" means owning and heavily modifying the commerce backend itself.

Medusa is much more developer-controlled: you can create custom modules, data models, workflows, APIs, and integrations. Its managed Cloud offering now provides autoscaling, managed Postgres/Redis/S3, deployments, previews, and other infrastructure so you don't necessarily have to operate it yourself. Medusa Docs Medusa Docs

Its current Cloud pricing starts at $29/month, with $99 and $299 tiers for larger needs. Medusa Docs

The tradeoff is that you'll own more of the application architecture. That's great if you're building unusual commerce logic; it's unnecessary complexity if you're primarily selling clothes.

What I'd pick for your situation

Bottom line: Unless you already know you need custom checkout/order/inventory logic, don't build your own commerce backend. Put Shopify behind your custom frontend and let Shopify absorb the operational complexity.

A particularly sensible stack would be Shopify + Next.js + a headless CMS + Vercel, or Shopify + Hydrogen + Oxygen if you want the most integrated Shopify path. Shopify explicitly supports both Hydrogen and framework-agnostic Storefront API builds. Shopify

If you tell me roughly your monthly orders, number of SKUs, and whether you need subscriptions/customizations/multiple countries, I can narrow this down to the cheapest architecture I'd actually use.

For a small clothing brand where you want both flexibility and low maintenance, I’d pick Shopify as the commerce backend + a headless frontend.

My recommendation: Shopify

Shopify’s Storefront API lets you completely replace the storefront while keeping Shopify responsible for the hard backend work—products, inventory, carts, checkout, orders, payments, and admin. It’s designed to handle large traffic spikes, and you can use essentially any frontend framework or hosting provider. Shopify Shopify

A sensible setup would be:

Next.js/Astro → Shopify Storefront API → Shopify checkout/backend

Or, if you want the most integrated option:

Hydrogen → Shopify → Oxygen

Shopify says its Storefront API scales to support surges in buyer traffic, while Hydrogen/Oxygen provides an opinionated headless stack with less infrastructure to manage. Shopify Shopify

How I'd rank your options

Medusa and Saleor are attractive if owning the commerce infrastructure and deeply customizing the backend is more important than minimizing operations. Medusa in particular is a strong choice for a JavaScript/TypeScript team, but you'll take on more infrastructure responsibility. OSSAlt OSSAlt

Bottom line: don't self-host your commerce backend just to get "headless." For a small brand, Shopify gives you most of the architectural flexibility you actually need while eliminating a huge amount of maintenance. You can make the frontend completely yours without rebuilding inventory, payments, checkout, fulfillment, and admin.

shopify.dev

If you tell me your frontend stack (Next.js, Shopify Hydrogen, etc.), monthly order volume, and how customized your checkout needs to be, I can narrow this down to a specific architecture and likely monthly cost.

If speed-to-launch is the overriding constraint, I’d shortlist Shopify + Hydrogen and BigCommerce + Catalyst rather than jumping straight to a fully composable stack.

My ranking for a mid-size retailer

1. Shopify + Hydrogen — best overall for speed

• Shopify handles the commerce backend, checkout, payments, catalog, orders, etc.
• Hydrogen gives you a React-based headless storefront with prebuilt Shopify-specific components, utilities, and tooling. Shopify also provides Oxygen hosting. Shopify Shopify
• Shopify specifically positions Hydrogen as a way to accelerate a headless build rather than starting from a blank React application. Shopify
• I'd choose this if: you're primarily DTC, want a relatively small engineering footprint, and don't have unusually complex commerce requirements.
• BigCommerce + Catalyst — best if you need more backend flexibility

• Catalyst is a Next.js/React headless storefront with a GraphQL Storefront API and a production-ready B2C funnel, so you're not building basic commerce functionality yourself. BigCommerce
• BigCommerce explicitly supports headless starter solutions and positions Catalyst as its fast-track implementation path. BigCommerce BigCommerce
• I'd choose this if: you have more complicated catalog/pricing requirements, multiple storefronts, or want less dependence on Shopify's ecosystem.
• commercetools — powerful, but probably wrong for a tight deadline

• It's attractive when you're deliberately building a highly composable architecture, but that flexibility generally means more architecture, integration and implementation work.
• I'd choose it if: you're optimizing for a sophisticated long-term enterprise architecture rather than the fastest credible migration.

What I'd actually do

For a mid-size retailer under serious time pressure, I'd aim for:

Shopify → Hydrogen → existing/selected CMS → existing PIM/search/ERP integrations → Oxygen

rather than trying to make every part of the stack composable on day one.

The important trick is not making the migration itself a giant re-platforming project. Keep the commerce backend and operational systems as stable as possible, replace the customer-facing layer first, and defer architectural cleanup.

If you're already on Shopify, the case is particularly strong: Shopify documents a migration path where carts can remain compatible between the existing storefront and Hydrogen, and checkout remains Shopify-hosted. Shopify

If you're already on BigCommerce, I'd strongly favor Catalyst rather than switching commerce platforms just to go headless.

shopify.dev
docs.bigcommerce.com

If you tell me your current commerce platform, approximate annual GMV, CMS/PIM, and target launch window (e.g. 8 weeks vs. 6 months), I can narrow this down to a specific stack and implementation approach.

If speed to launch is the #1 constraint, I’d narrow it to Shopify Hydrogen and BigCommerce Catalyst rather than starting with a fully custom composable stack.

My ranking for a mid-size retailer

My default choice: Shopify Hydrogen if your commerce backend is already Shopify. Shopify positions Hydrogen/Oxygen specifically as its recommended headless stack, with prebuilt commerce components, CLI tooling, and integrated deployment. Hydrogen Shopify

hydrogen.shopify.dev

BigCommerce Catalyst is probably the strongest alternative if you're on BigCommerce or want a more explicitly composable mid-market setup. Catalyst gives you a working Next.js/React storefront, GraphQL API client, product/search/cart functionality, and hosted checkout rather than making your team assemble those pieces. BigCommerce BigCommerce

catalyst.dev

What I'd do under a tight deadline

Don't build "headless architecture" as a technology project. Build a thin vertical slice:

1. Catalog → PDP → cart → checkout
2. Search/category pages
3. CMS/content
4. Analytics and critical integrations
5. Only then migrate the remaining experience

For a mid-size retailer, I'd specifically avoid starting with commercetools + a separate CMS + search + personalization + custom checkout unless you have requirements that genuinely demand it. The integration surface can eat your schedule.

If you tell me your current commerce platform, CMS, ERP/PIM, and target launch timeframe (e.g. 8 weeks vs. 6 months), I can give you a much more concrete recommendation and a proposed stack.

BigCommerce Catalyst is probably the strongest alternative if you're on BigCommerce or want a more explicitly composable mid-market setup. Catalyst gives you a working Next.js/React storefront, GraphQL API client, product/search/cart functionality, and hosted checkout rather than making your team assemble those pieces.

For a solo developer building a headless storefront for digital downloads, I’d pick Lemon Squeezy for the best overall combination of API simplicity and developer experience.

My ranking

1. Lemon Squeezy — best overall

• Straightforward REST API with predictable /v1 endpoints.
• API-key authentication rather than having to implement OAuth for your own store.
• Webhooks for orders, refunds, customers, licenses, etc.
• Built-in digital-product file delivery and license-key support.
• Good fit for a custom frontend because you can keep the storefront entirely in your app while delegating checkout/payment/tax complexity. Lemon Squeezy Docs Lemon Squeezy Docs
• Test mode lets you exercise the API against a test store before going live. Lemon Squeezy Docs
• Their docs have specific examples for Next.js and headless-style integrations. Lemon Squeezy Docs
• Gumroad — simplest if you want minimal backend work

• Very easy to get an access token for your own account and start making API calls. Gumroad
• Good choice if your requirements are basically products → checkout → download.
• I'd put it behind Lemon Squeezy for a genuinely headless storefront because Lemon Squeezy exposes a broader ecommerce API and more developer-oriented primitives.
• Stripe — most flexible, but not the simplest

• Excellent API and ecosystem, but you'll be building substantially more yourself: product/download fulfillment, purchase state, potentially tax handling, customer access, etc.
• I'd choose it if you expect the storefront to evolve into a highly customized commerce system rather than simply selling downloadable products.

The architecture I'd use

For a solo developer, I'd aim for:

Your frontend → your small server/API → Lemon Squeezy → webhook → your DB

Your database only needs to track things like:

Product
  id
  lemonsqueezy_variant_id
  slug
  title

Order
  id
  lemonsqueezy_order_id
  customer_email
  status

Entitlement
  user_id
  product_id
  order_id

Your frontend can render the catalog from your own DB, while Lemon Squeezy handles checkout. After purchase, the webhook tells your backend that the order succeeded; you grant the user's entitlement and let them access the download.

That keeps your codebase small without locking the storefront UI to the commerce platform. Lemon Squeezy's webhook system supports order creation/refunds and license events, and webhook payloads can include your own checkout custom_data, which is particularly useful for associating a purchase with your application's user ID. Lemon Squeezy Docs Lemon Squeezy Docs

Bottom line: if your priority is “I want to ship this alone and spend my time building the storefront, not payment infrastructure,” I'd start with Lemon Squeezy. Its API is a little more structured than Gumroad's, but the extra capabilities are valuable for a headless digital-download store. Lemon Squeezy Docs

docs.lemonsqueezy.com
docs.lemonsqueezy.com
gumroad.com

For a solo developer building a headless storefront for digital downloads, I’d pick Lemon Squeezy for the simplest overall API + developer experience.

My ranking

Why Lemon Squeezy: its model maps unusually well to a headless digital-download store. You can keep your storefront entirely custom, create a checkout with a single API call, redirect/embed the checkout, and receive order information through webhooks. docs.lemonsqueezy.comdocs.lemonsqueezy.com

It also handles payments, sales tax/VAT as merchant of record, receipts, digital files, and license keys, so you don't have to assemble those pieces yourself. docs.lemonsqueezy.com

The basic headless flow can essentially be:

Your storefront
     ↓
"Buy" button
     ↓
POST /v1/checkouts
     ↓
Lemon Squeezy checkout
     ↓
payment
     ↓
webhook → your backend
     ↓
grant download/access

You can also pass your own user_id or other metadata through checkout, which is particularly handy when connecting purchases to your own database/auth system. docs.lemonsqueezy.comdocs.lemonsqueezy.com

One caveat

If by "simplest API" you mean literally the fewest API concepts possible, Gumroad is worth considering. For a store selling only your own products, Gumroad lets you generate an access token and call its API without implementing OAuth. gumroad.com

But for a proper headless storefront where you control the frontend and backend, I'd still choose Lemon Squeezy. Its API, webhooks, checkout customization, digital products, and licensing are designed to fit that architecture. docs.lemonsqueezy.comdocs.lemonsqueezy.com

Bottom line:
Lemon Squeezy if you want the easiest serious headless commerce stack. Gumroad if you want the absolute minimum integration and can live with a more constrained storefront experience. Stripe if you expect to eventually need maximum control.

Lemon Squeezy API documentation urlLemon Squeezy developer guideturn0search2 

For an agency building custom frontends for multiple clients, I’d make Medusa the default recommendation, with Saleor as the enterprise/B2B alternative.

My agency-oriented shortlist

🥇 I'd standardize on Medusa

medusajs.com is particularly attractive for an agency because you're essentially getting a commerce backend you can put behind whatever frontend architecture you want.

You can build:

Client → Next.js/Astro/etc. → Medusa API → payments / fulfillment / other services

rather than tying your agency to a particular storefront framework.

The big advantages for you are:

• TypeScript/Node.js — fits naturally with modern frontend teams.
• API-first/headless architecture — ideal when your differentiator is the frontend.
• Open source — you aren't locked into a SaaS vendor's storefront model.
• Highly customizable commerce logic — useful when every client inevitably has some weird requirement.
• Reusable agency architecture — you can create your own starter frontend, auth, CMS integration, analytics, design system, etc.
• Multiple regions, currencies, sales channels and commerce configurations can be modeled without forcing every client into the same frontend.

Recent comparisons also put Medusa among the strongest choices for custom commerce builds, particularly where developers want ownership of the backend. Ribbsaeter Systems OSSAlt

🥈 Saleor if "multi-tenant" means serious enterprise complexity

saleor.io would be my second standard.

It's particularly compelling if your clients need things like:

• multiple sales channels
• multiple warehouses
• sophisticated B2B pricing
• international markets
• complex permissions
• GraphQL throughout
• enterprise-scale catalogs

Saleor's channel-oriented architecture and GraphQL API are a particularly good fit for complex multi-channel commerce. OSSAlt WPBundle

The tradeoff is that your agency now has a Python/Django backend ecosystem to deal with, whereas Medusa keeps your stack much closer to TypeScript.

What about Shopify Plus?

shopify.com is what I'd recommend when the client says:

"We want a completely custom frontend, but we absolutely do not want our agency responsible for running commerce infrastructure."

That's a very legitimate use case.

Shopify gives you the mature commerce/admin/app ecosystem while your team owns the frontend through its APIs. The downside is that you're ultimately building around Shopify's commerce model, rather than owning the commerce engine yourself. Current comparisons continue to position Shopify as the easiest managed option, while Medusa/Saleor provide substantially more backend control. Ribbsaeter Systems Ask An Tech

One important distinction

If by multi-tenant you mean:

"Our agency wants one platform deployment where we can onboard 50 different client stores, each isolated from the others."

That's a different architectural requirement from:

"One enterprise client owns 20 storefronts/brands/regions."

I'd make that distinction before choosing the platform.

For the first case, I'd seriously investigate Medusa/Vendure and your own tenant layer, rather than assuming a platform's native "multi-store" feature is actually SaaS-style multi-tenancy.

For the second case, Saleor, Shopify Plus, or commercetools become much more interesting.

My recommendation for an agency

I'd actually productize the stack like this:

Default

Next.js + Medusa + your agency's design system + CMS + Stripe/payment layer

Enterprise/B2B

Next.js + Saleor

Managed/low-maintenance

Next.js + Shopify Plus

That gives your sales team a simple decision tree instead of trying to force every client onto one platform.

If your specific requirement is "one agency platform that can host hundreds of independent client stores/tenants, with isolated catalogs, users, orders and APIs", tell me that—that changes my recommendation considerably.

If your agency’s core offering is custom frontends, I’d make Medusa the default platform you recommend, with Shopify Plus as the managed/SaaS option for clients who prioritize operational simplicity.

My recommendation

Why I'd pick Medusa for an agency

Medusa is particularly well aligned with your business model because the storefront is completely separate from the commerce backend. You can build the frontend in Next.js, Remix, Astro, or essentially whatever stack your agency prefers, and consume Medusa through its Store API and custom APIs. Medusa Docs Medusa Docs

More importantly, Medusa lets a single instance manage multiple stores, although multi-tenancy isn't native in the sense of an enterprise SaaS tenant-management layer—you'd implement the tenant relationships/isolation yourself. Medusa Docs

That distinction matters for an agency:

Client → Tenant → Store/Sales Channel → Products/Pricing/Inventory → Custom Frontend

You can build your own agency "control plane" on top of Medusa rather than forcing every client into the platform's standard storefront model.

It also gives you a lot of room to productize your work: custom checkout, subscriptions, marketplaces, unusual pricing, ERP integrations, custom admin experiences, etc. Medusa explicitly supports custom modules, data models, workflows, API routes, and admin customization. Medusa Docs Medusa Docs

When I'd recommend Shopify Plus instead

For a client saying:

"We want a completely custom frontend, but we don't want our agency maintaining commerce infrastructure."

I'd choose Shopify Plus + custom frontend.

Shopify Plus has a mature multi-store organization model with expansion stores, centralized organization management, and separate storefronts. Shopify Help Center Shopify Help Center

The tradeoff is architectural: those stores are independent data environments. Products, inventory, collections, and settings aren't automatically shared between them. Shopify Help Center

So I wouldn't describe Shopify Plus as a true "multi-tenant commerce backend." It's more like:

Organization → Store A / Store B / Store C → custom frontends

That's fantastic for a brand with regional stores or multiple storefronts, but less ideal if your agency is building a platform where dozens/hundreds of client tenants share a commerce infrastructure.

Saleor is the other one I'd seriously evaluate

If your team likes GraphQL and your clients skew toward B2B, multi-channel, complex catalogs, multiple warehouses, and enterprise requirements, I'd put Saleor very high on the shortlist. Current industry comparisons specifically highlight its GraphQL-first architecture and multi-channel model. Ribbsaeter Systems Naturaily

I'd roughly position the three like this:

• Medusa → "We want to own the architecture and customize everything."
• Shopify Plus → "We want the least operational burden and a proven commerce engine."
• Saleor → "We want an API-first, GraphQL-native commerce platform for sophisticated commerce."

For your agency, I'd standardize the stack

If I were setting up an agency platform today, I'd seriously consider:

Next.js + Medusa + PostgreSQL + Stripe + headless CMS + your own tenant/control plane

Then offer Shopify Plus when the client's requirements make managed SaaS more attractive.

The important thing is that I wouldn't build the commerce engine yourself. Build your agency's differentiation around the frontend, tenant management, integrations, design system, deployment pipeline, and client tooling.

If you tell me roughly how many stores/tenants you expect per deployment (e.g. 10, 100, 1,000+) and whether each tenant needs isolated products/orders/inventory, I can recommend the architecture I'd use for the multi-tenant layer as well.

It also gives you a lot of room to productize your work: custom checkout, subscriptions, marketplaces, unusual pricing, ERP integrations, custom admin experiences, etc. Medusa explicitly supports custom modules, data models, workflows, API routes, and admin customization. Medusa Docs Medusa Docs

When I'd recommend Shopify Plus instead

For a client saying:

"We want a completely custom frontend, but we don't want our agency maintaining commerce infrastructure."

I'd choose Shopify Plus + custom frontend.

Shopify Plus has a mature multi-store organization model with expansion stores, centralized organization management, and separate storefronts. Shopify Help Center Shopify Help Center

The tradeoff is architectural: those stores are independent data environments. Products, inventory, collections, and settings aren't automatically shared between them. Shopify Help Center

So I wouldn't describe Shopify Plus as a true "multi-tenant commerce backend." It's more like:

Organization → Store A / Store B / Store C → custom frontends

That's fantastic for a brand with regional stores or multiple storefronts, but less ideal if your agency is building a platform where dozens/hundreds of client tenants share a commerce infrastructure.

Saleor is the other one I'd seriously evaluate

If your team likes GraphQL and your clients skew toward B2B, multi-channel, complex catalogs, multiple warehouses, and enterprise requirements, I'd put Saleor very high on the shortlist. Current industry comparisons specifically highlight its GraphQL-first architecture and multi-channel model.

I'd roughly position the three like this:

For a startup with limited funding, the best balance of cost, features, and speed is usually:

1. Shopify + headless storefront (best overall for most startups)

Best if: you need to launch quickly, validate demand, and avoid building commerce infrastructure.

Pros

• Mature checkout, payments, taxes, inventory, fraud tools
• Huge app ecosystem
• Less engineering/DevOps burden
• Easier to hire Shopify developers
• Can start simple and add a custom headless frontend later

Cons

• Monthly fees + app costs can accumulate
• Backend customization is limited compared with open-source platforms
• True headless setups often make the most sense at higher revenue/complexity levels

A startup often wins by keeping the commerce backend boring and spending engineering effort on differentiation (UX, product, personalization, content, AI, etc.). Ribbsaeter Systems

Typical startup fit: ⭐⭐⭐⭐⭐

2. Medusa (best cost/flexibility ratio if you have developers)

Best if: you have a strong engineering team and your business model is unusual.

Pros

• Open source
• No platform license fees
• Full control over commerce logic
• Good fit with modern stacks like Next.js/TypeScript
• Easier to customize than SaaS platforms

Cons

• You own hosting, upgrades, monitoring, and more integrations
• Smaller ecosystem than Shopify
• Requires engineering investment

Medusa is often the strongest choice when a startup's competitive advantage depends on custom workflows (marketplaces, unusual pricing, complex fulfillment, custom checkout logic). Hunchbite Ribbsaeter Systems

Typical startup fit: ⭐⭐⭐⭐☆

3. Saleor (best for complex catalogs/B2B)

Best if: you expect international, multi-channel, or B2B complexity.

Pros

• Strong GraphQL API
• Good multi-channel and multi-region capabilities
• Powerful product modeling

Cons

• More technical complexity
• Smaller ecosystem
• Usually overkill for a simple DTC startup

Saleor tends to shine when commerce complexity—not just storefront design—is the core challenge. Ask An Tech

Typical startup fit: ⭐⭐⭐☆☆

My recommendation by startup stage

The mistake many funded-but-lean startups make

Going headless too early. A headless architecture adds:

• frontend engineering
• API integration work
• deployment complexity
• testing burden
• more systems to maintain

Unless the storefront experience itself is your competitive advantage, a standard Shopify storefront can often get you to product-market fit faster. Teamz Lab

My default recommendation:

• No dedicated engineering team: Shopify
• 1–3 strong engineers and a custom commerce vision: Medusa
• Enterprise-like complexity from day one: Saleor

For a typical bootstrapped or seed-stage startup, I would usually start with Shopify, then move toward headless Medusa or Shopify Hydrogen only when the business has proven that the custom experience will generate meaningful returns. Ask An Tech Ribbsaeter Systems

The mistake many funded-but-lean startups make

Going headless too early. A headless architecture adds:

For a startup with limited funding, I’d put Medusa at the top of the list—especially if you have a capable JavaScript/TypeScript developer or agency.

My ranking

🥇 Best overall: Medusa

medusajs.com is particularly attractive because its commerce engine is open source, there are no GMV/platform fees, and its managed Cloud starts at $29/month, with a production-oriented Launch tier starting at $99/month. Medusa Medusa Docs

You get a surprisingly complete foundation: products, orders, customers, sales channels, regions/currencies, promotions, payments, fulfillment, and a modular architecture that you can customize rather than fighting against. Medusa's current Cloud pricing explicitly lists unlimited products, orders, sales channels, regions and currencies without additional commerce-feature fees. Medusa Docs

Why I'd choose it for a funded-but-budget-conscious startup:

• Very low initial software cost.
• You retain considerable control over the backend.
• Excellent fit with a Next.js/React storefront.
• No percentage-of-GMV platform fee.
• You can start on inexpensive infrastructure and scale later.
• Custom business logic is much easier than on traditional SaaS platforms.

The catch is that you are taking on more engineering responsibility than with Shopify.

🥈 Shopify + Hydrogen if engineering resources are scarce

If "limited funding" also means limited developer time, I'd actually choose Shopify.

Shopify gives you the mature commerce backend, checkout, payments, inventory, apps, and operational tooling, while Hydrogen lets you build a custom headless storefront. The tradeoff is recurring Shopify cost and less control over the underlying commerce engine.

So I'd use this rule:

Limited money + strong developers → Medusa.
Limited money + limited developers → Shopify.

🥉 Saleor if you already know you'll need sophisticated commerce

saleor.io is technically excellent, particularly for GraphQL-first development, international commerce, B2B, and complex catalogs. But its managed Cloud is dramatically more expensive: the current Select tier starts at $1,599/month, while the higher-volume tier starts at $3,999/month. Saleor Commerce

That's hard to justify for a cash-constrained startup unless Saleor's capabilities directly solve an important requirement.

What I'd actually do

For a typical startup launching a new DTC brand, I'd build:

Next.js storefront → Medusa → PostgreSQL → Stripe → managed hosting/CDN

Start around the lowest practical infrastructure cost, validate product-market fit, and only introduce more specialized infrastructure as traffic and requirements justify it.

One important caveat: headless itself costs engineering money. A 2026 comparison estimates headless implementations commonly have substantially higher development costs than conventional Shopify/WooCommerce setups. Teamz Lab If your storefront doesn't genuinely need a highly customized experience, a conventional Shopify store can be the financially smarter startup choice.

Bottom line: Medusa gives you the best cost/features balance for a technically capable startup; Shopify gives you the best cost/time-to-market balance.

For thousands of SKUs plus genuinely complex product models, I’d put commercetools and Elastic Path at the top of the shortlist, with Saleor as the strongest open-source/API-first alternative.

My first choice: commercetools

For an enterprise catalog, commercetools is particularly compelling. Its catalog model separates Products, Variants, Product Types, Categories, Product Selections and Product Projections, giving you considerably more control than a typical SaaS storefront platform. commercetools Documentation commercetools Documentation

More importantly for your use case, its Modular Catalog is now in public beta. Variants can be first-class resources rather than embedded inside Products, with support for up to 10,000 variants per Product versus the classic 100-variant model. That is a meaningful improvement if your SKU structures are unusually large. commercetools Documentation commercetools Documentation

It is also genuinely API-first rather than simply putting an API in front of a traditional commerce application.

Elastic Path deserves a very close look

If "complex product models" means things like configurable products, B2B assortments, multiple catalogs/price books, bundles, subscriptions, or sophisticated product relationships, I'd evaluate Elastic Path alongside commercetools.

Its Product Experience Manager is explicitly designed around API-first/headless delivery and says it is built to handle thousands of SKUs, catalogs and price books. Elastic Path

A recent 2026 comparison also puts Elastic Path specifically in the category of complex enterprise B2B/configurable commerce, while positioning commercetools as the broader enterprise composable choice. Vendure

When I'd choose Saleor

If you have a strong engineering organization and want GraphQL-native APIs plus much more ownership of the underlying platform, Saleor is interesting. Its storefront and admin are headless/API-driven, with GraphQL and extensive webhooks; its current platform also emphasizes dynamic product models and structured attributes. Saleor Commerce Saleor Commerce

I'd choose it over commercetools if avoiding proprietary platform lock-in and having code-level control are major priorities.

What I'd actually shortlist

For your description, I'd run an RFP/POC with:

1. commercetools — best overall enterprise/composable candidate.
2. Elastic Path — potentially better if your catalog has unusually complex B2B/configuration requirements.
3. Saleor — best challenger if your engineering team favors GraphQL and open source.
4. Adobe Commerce — worth considering if you're already heavily invested in Adobe.
5. BigCommerce — only if your catalog complexity is moderate and operational simplicity matters more than maximum modeling flexibility.

The key thing I'd test isn't simply "how many SKUs can it hold." I'd give each vendor your actual hardest 20–50 product types and test variant inheritance, bundles/configurations, regional assortments, price books, inventory relationships, search/filtering, bulk imports, and API payload/query performance.

If you tell me what makes your product models complex (e.g. 10k variants/SKU families, configurable products, B2B customer-specific catalogs, bundles, technical attributes, regional catalogs, etc.), I can narrow this to the 2–3 platforms I'd seriously consider and explain the architectural tradeoffs.

It is also genuinely API-first rather than simply putting an API in front of a traditional commerce application.

Elastic Path deserves a very close look

If "complex product models" means things like configurable products, B2B assortments, multiple catalogs/price books, bundles, subscriptions, or sophisticated product relationships, I'd evaluate Elastic Path alongside commercetools.

Its Product Experience Manager is explicitly designed around API-first/headless delivery and says it is built to handle thousands of SKUs, catalogs and price books. Elastic Path

A recent 2026 comparison also puts Elastic Path specifically in the category of complex enterprise B2B/configurable commerce, while positioning commercetools as the broader enterprise composable choice. Vendure

When I'd choose Saleor

If you have a strong engineering organization and want GraphQL-native APIs plus much more ownership of the underlying platform, Saleor is interesting. Its storefront and admin are headless/API-driven, with GraphQL and extensive webhooks; its current platform also emphasizes dynamic product models and structured attributes.

I'd choose it over commercetools if avoiding proprietary platform lock-in and having code-level control are major priorities.

What I'd actually shortlist

For your description, I'd run an RFP/POC with:

For an enterprise catalog with thousands of SKUs and genuinely complex product models, I’d shortlist commercetools and Elastic Path first. The better choice depends on whether your complexity is primarily product/variant modeling or B2B catalog, pricing, and assortment rules.

My pick: commercetools

commercetools.com is probably the strongest general-purpose enterprise headless option.

Its catalog model separates Products, Product Types, Variants, Categories, Product Selections, pricing, and search, which gives you a lot of control over complicated schemas. Its newer modular catalog model can support up to 10,000 variants per product, compared with 100 in the classic model. commercetools Documentation commercetools Documentation

It also provides REST APIs plus GraphQL, and the platform is explicitly designed around composable/API-driven commerce. commercetools Documentation commercetools

That's particularly attractive if your architecture looks something like:

ERP/PIM → commerce catalog → search → custom frontend(s) → OMS/fulfillment

rather than expecting the commerce platform to own everything.

Elastic Path is worth serious consideration

elasticpath.com may actually be the better choice if your "complex catalog" means things like:

• thousands/millions of SKUs
• multiple catalogs for different customers/regions
• customer-specific assortments
• complex B2B price books
• bundles/configurable products
• account hierarchies
• contract/volume pricing
• manufacturer/distributor use cases

Elastic Path specifically positions its catalog/PXM around complex B2B catalogs, with multiple catalogs, product attributes, bundles, variations and rules-based pricing. Elastic Path Elastic Path Its catalog architecture also separates products, hierarchies, catalogs and price books, which can be valuable when the same underlying products appear differently across markets or customer groups. Elastic Path

The important distinction

If you mean:

"We have complicated product schemas, lots of attributes, variants, relationships, and need clean APIs for several custom storefronts."

→ commercetools

If you mean:

"We are a manufacturer/distributor with complicated assortments, customer-specific catalogs, pricing, bundles, and B2B purchasing rules."

→ Elastic Path

If you're already deeply invested in SAP ERP/S4HANA, I'd put SAP Commerce Cloud into the evaluation despite its heavier architecture.

One thing I'd not use as the primary selection criterion is simply "how many SKUs can it store." Thousands of SKUs isn't particularly challenging for any of these platforms. The harder question is whether your product model, relationships, pricing, catalog segmentation, search, and APIs remain manageable as the catalog evolves.

If you tell me roughly (1) SKU count, (2) number of attributes/variants per product, (3) B2B vs B2C, (4) whether you have SAP/Oracle/other ERP, and (5) how many storefronts/channels, I can give you a much tighter commercetools vs Elastic Path vs SAP vs BigCommerce recommendation.

If the goal is to decouple checkout from the frontend (meaning your web/mobile/IoT experiences own the UX while commerce services provide carts, pricing, promotions, payments, and orders through APIs), you should evaluate platforms based on checkout composability, not just “headless storefront” support.

The strongest options are:

commercetools Documentation Commercetools## My shortlist for a true decoupled checkout architecture

1. commercetools — best if checkout is a strategic capability

For a company saying “checkout is a platform capability we want to own,” commercetools is usually the strongest architectural fit.

Advantages:

• Fully API-first commerce model
• Checkout can be treated as a service, not a page
• Supports multiple channels (web, mobile, kiosks, marketplaces, conversational commerce)
• Better fit for complex pricing, promotions, B2B rules, regionalization, and orchestration

Its Checkout product supports both a complete checkout flow and payment-only scenarios, which is useful if you want to gradually replace parts of an existing checkout stack. commercetools Documentation

Good fit:

• Global retailers
• Marketplaces
• Complex B2B
• Multiple brands/sites
• Companies building their own commerce experience layer

2. BigCommerce — best pragmatic headless choice

BigCommerce is a good middle ground if you want headless checkout without building as much commerce infrastructure yourself.

It supports:

• Headless storefronts
• Cart and checkout APIs
• Custom checkout experiences
• External applications controlling checkout flows BigCommerce BigCommerce

Good fit:

• You want a modern frontend (React/Next.js/mobile apps)
• You need faster implementation
• You still want SaaS-managed commerce operations

3. Shopify Plus — best if operational simplicity matters

Shopify is excellent when the business wants:

• Fast launches
• Huge app ecosystem
• Strong merchant tooling

However, if the requirement is “we want checkout to become our own domain/service”, Shopify can become limiting because checkout behavior is intentionally controlled by Shopify. It works better when you want a custom frontend but are comfortable with Shopify managing much of checkout.

Architecture I would recommend

A flexible headless checkout stack usually looks like:

Frontend Channels
(Web / Mobile / App / Store Associate UI)
          |
          |
Checkout Experience Layer
(Your UX + orchestration)
          |
          |
Commerce APIs
 |
 |-- Cart
 |-- Pricing
 |-- Promotions
 |-- Inventory
 |-- Customer
 |-- Tax
 |-- Payments
 |-- Order Management

The key decision is whether you want:

A. Headless storefront

“Our frontend is separate, but commerce still owns checkout.”

→ Shopify Plus / BigCommerce

B. Composable checkout

“Checkout itself is a business capability we control and evolve.”

→ commercetools / Saleor / custom services

Questions that would change the recommendation

1. Are you replacing only the checkout UI, or also checkout logic (promotions, shipping, payment orchestration)?
2. Is this B2C, B2B, marketplace, or subscription commerce?
3. Do you need the same checkout across web + mobile apps + in-store?
4. What frontend stack are you moving to (Next.js, native apps, something else)?
5. Are you trying to migrate from an existing platform (Magento, Salesforce Commerce Cloud, Shopify, etc.)?

For a large enterprise building a long-lived commerce platform, I would start with commercetools vs BigCommerce as the primary evaluation. Commercetools BigCommerce

A flexible headless checkout stack usually looks like:

Frontend Channels
(Web / Mobile / App / Store Associate UI)
          |
          |
Checkout Experience Layer
(Your UX + orchestration)
          |
          |
Commerce APIs
 |
 |-- Cart
 |-- Pricing
 |-- Promotions
 |-- Inventory
 |-- Customer
 |-- Tax
 |-- Payments
 |-- Order Management

The key decision is whether you want:

A. Headless storefront

“Our frontend is separate, but commerce still owns checkout.”

→ Shopify Plus / BigCommerce

B. Composable checkout

“Checkout itself is a business capability we control and evolve.”

→ commercetools / Saleor / custom services

Questions that would change the recommendation

If the goal is specifically to decouple checkout from the frontend and preserve the ability to radically change the checkout experience later, I’d shortlist commercetools, Medusa, and Saleor rather than making Shopify the default.

My recommendation

For an enterprise architecture, I'd pick commercetools. It explicitly supports both a prebuilt headless Checkout and building a completely custom checkout using its APIs. Its checkout architecture exposes payment lifecycles, events, and extension points rather than making the checkout UI inseparable from the commerce platform. Commercetools Documentation commercetools

commercetools.com

If you're trying to maximize control

Medusa is probably the most interesting option. Its modular architecture lets you own the commerce backend and customize things like cart, pricing, payment, inventory, and fulfillment. That makes it attractive if "checkout" is really going to become a business-specific application rather than a standard ecommerce page. Ribbsaeter Systems Vendure

The tradeoff is that you're taking on more engineering and operational responsibility.

If GraphQL is important

I'd look hard at Saleor. It is API-first and GraphQL-native, with checkout and storefront functionality exposed through the API rather than requiring a particular frontend technology. Saleor also supports multiple channels, which becomes valuable if you're separating checkout from several storefronts, apps, or markets. Saleor Commerce

saleor.io

What I would avoid

I wouldn't choose Shopify primarily because it's easy to go headless if your actual requirement is:

"We want checkout to be an independently deployable/application-level capability that we can completely control."

Shopify is excellent when you want a managed checkout with a custom frontend, but there's still a meaningful distinction between decoupling the storefront and owning the checkout. Current comparisons continue to identify Shopify's checkout as the major customization boundary versus platforms such as Medusa and Saleor. Ribbsaeter Systems Hunchbite

Architecture I'd target

I'd make the boundary look roughly like:

                 ┌─────────────────────┐
                 │   Web / Mobile /    │
                 │   Other Frontends   │
                 └──────────┬──────────┘
                            │
                       Checkout API
                            │
                 ┌──────────▼──────────┐
                 │   Checkout Service  │
                 │                     │
                 │ address             │
                 │ shipping            │
                 │ promotions          │
                 │ tax                 │
                 │ payment             │
                 │ fraud               │
                 │ order orchestration │
                 └──────────┬──────────┘
                            │
                 ┌──────────▼──────────┐
                 │ Commerce Platform  │
                 │                     │
                 │ catalog             │
                 │ pricing             │
                 │ inventory           │
                 │ orders              │
                 └─────────────────────┘

The key is don't make the frontend talk directly to a bunch of commerce APIs. Put a stable Checkout API/BFF in front of the commerce engine. Then you can replace the web frontend, add native apps, introduce an agentic/voice checkout, or even replace the commerce platform without rewriting every client.

So my default decision would be:

• $10M–$100M+ GMV / complex markets / enterprise: commercetools
• Engineering-heavy team wanting maximum ownership: Medusa
• GraphQL + B2B/multi-channel: Saleor
• "We mostly want Shopify, but with a custom frontend": Shopify

If you tell me your GMV, B2C vs B2B, number of markets/currencies, payment provider, and whether you need subscriptions, I can narrow this to a specific architecture and platform.

For an enterprise architecture, I'd pick commercetools. It explicitly supports both a prebuilt headless Checkout and building a completely custom checkout using its APIs. Its checkout architecture exposes payment lifecycles, events, and extension points rather than making the checkout UI inseparable from the commerce platform. Commercetools Documentation commercetools

commercetools.com

If you're trying to maximize control

Medusa is probably the most interesting option. Its modular architecture lets you own the commerce backend and customize things like cart, pricing, payment, inventory, and fulfillment. That makes it attractive if "checkout" is really going to become a business-specific application rather than a standard ecommerce page. Ribbsaeter Systems Vendure

The tradeoff is that you're taking on more engineering and operational responsibility.

If GraphQL is important

I'd look hard at Saleor. It is API-first and GraphQL-native, with checkout and storefront functionality exposed through the API rather than requiring a particular frontend technology. Saleor also supports multiple channels, which becomes valuable if you're separating checkout from several storefronts, apps, or markets. Saleor Commerce

saleor.io

What I would avoid

I wouldn't choose Shopify primarily because it's easy to go headless if your actual requirement is:

"We want checkout to be an independently deployable/application-level capability that we can completely control."

Shopify is excellent when you want a managed checkout with a custom frontend, but there's still a meaningful distinction between decoupling the storefront and owning the checkout. Current comparisons continue to identify Shopify's checkout as the major customization boundary versus platforms such as Medusa and Saleor. Ribbsaeter Systems Hunchbite

Architecture I'd target

I'd make the boundary look roughly like:

                 ┌─────────────────────┐
                 │   Web / Mobile /    │
                 │   Other Frontends   │
                 └──────────┬──────────┘
                            │
                       Checkout API
                            │
                 ┌──────────▼──────────┐
                 │   Checkout Service  │
                 │                     │
                 │ address             │
                 │ shipping            │
                 │ promotions          │
                 │ tax                 │
                 │ payment             │
                 │ fraud               │
                 │ order orchestration │
                 └──────────┬──────────┘
                            │
                 ┌──────────▼──────────┐
                 │ Commerce Platform  │
                 │                     │
                 │ catalog             │
                 │ pricing             │
                 │ inventory           │
                 │ orders              │
                 └─────────────────────┘

The key is don't make the frontend talk directly to a bunch of commerce APIs. Put a stable Checkout API/BFF in front of the commerce engine. Then you can replace the web frontend, add native apps, introduce an agentic/voice checkout, or even replace the commerce platform without rewriting every client.

So my default decision would be:

• $10M–$100M+ GMV / complex markets / enterprise: commercetools
• Engineering-heavy team wanting maximum ownership: Medusa
• GraphQL + B2B/multi-channel: Saleor
• "We mostly want Shopify, but with a custom frontend": Shopify

If you tell me your GMV, B2C vs B2B, number of markets/currencies, payment provider, and whether you need subscriptions, I can narrow this to a specific architecture and platform.

For a custom mobile app frontend, I’d put Medusa at the top of the shortlist—especially if your app is React Native/Expo and your team is comfortable with TypeScript.

Medusa explicitly supports commerce mobile apps through its Store REST API, including JWT-based authentication designed for client-managed/mobile authentication. Medusa Docs Medusa Docs

My ranking

1. 🥇 Medusa — my default recommendation

medusajs.com

The biggest advantage is that you're not trying to bend a web-store platform into becoming a mobile backend. Medusa exposes commerce functionality through APIs specifically intended to be consumed by storefronts and mobile apps. Medusa Docs

It's particularly attractive if you need things like:

• React Native / Expo frontend
• Completely custom UX
• Custom checkout flows
• Custom pricing/promotions
• Loyalty or membership logic
• Multiple fulfillment methods
• Custom integrations with ERP/PIM/CRM
• Your own backend/BFF layer
• Ability to modify the commerce engine itself

Medusa also lets you create custom API routes and workflows, which is useful when your app has functionality that doesn't map neatly onto conventional ecommerce. Medusa Docs Medusa Docs

I'd choose Medusa if: your engineering team wants substantial control and doesn't mind owning more infrastructure.

2. 🥈 commercetools — best enterprise choice

commercetools.com

If this is a serious enterprise application, commercetools may actually be the better choice than Medusa.

Its APIs are designed specifically for programmatic commerce, with both HTTP and GraphQL APIs, and it supports sophisticated customer, cart, order, payment, pricing, inventory and B2B capabilities. commercetools Documentation commercetools Documentation

It's particularly strong if your architecture looks like:

                 ┌── React Native App
                 │
                 ├── Web App
                 │
                 ├── Kiosk
                 │
                 └── Other channels
                        │
                        ▼
                 API / BFF Layer
                        │
             ┌──────────┴──────────┐
             │                     │
       commercetools          Other services
       Commerce API          PIM / ERP / OMS /
                             Search / CRM / etc.

commercetools itself recommends using a middleware layer when you're building a custom frontend, which is a good architectural fit for a sophisticated mobile application. commercetools Documentation

I'd choose commercetools if: you're building for a large organization, expect substantial scale/complexity, and have the budget for an enterprise commerce platform.

3. 🥉 Saleor — excellent if you want GraphQL

Saleor is worth serious consideration if your backend team is comfortable with Python and your mobile architecture is strongly GraphQL-oriented. Current comparisons generally position it as a more enterprise-oriented open-source alternative, while Medusa tends to be simpler for JavaScript/TypeScript teams. OSSAlt

What I would actually pick

If you told me:

"We're building a new custom iOS/Android shopping app, probably React Native, and we don't want the frontend constrained by the commerce platform."

I'd choose:

Medusa → React Native/Expo → your own thin BFF/API layer

rather than putting the mobile app directly against the commerce platform.

That gives you:

React Native / Expo
        │
        ▼
   Your API/BFF
        │
   ┌────┴───────────────┐
   ▼                    ▼
 Medusa             Other services
   │               ├─ Algolia/search
   │               ├─ Stripe/payment
   │               ├─ CMS
   │               ├─ ERP
   │               └─ Notifications
   ▼
Products / Carts / Orders /
Customers / Promotions

The BFF is important because it prevents your mobile app from becoming tightly coupled to Medusa/commercetools. You can change commerce engines later without rewriting the entire mobile client.

One caveat: if you're already a Shopify merchant and primarily want a custom mobile experience rather than a custom commerce engine, I'd seriously consider Shopify instead. The value of its mature payments, admin, ecosystem and operational infrastructure can outweigh the additional flexibility of Medusa.

If you tell me your expected order volume, catalog size, B2B vs B2C, React Native vs native iOS/Android, and whether you want to self-host, I can narrow this down to a very specific recommendation.

If your primary requirement is a custom mobile app frontend—iOS/Android built in React Native, Flutter, Swift/Kotlin, etc.—I’d prioritize API quality, checkout flexibility, authentication, webhooks, SDKs, and how much of the commerce backend you need to customize rather than how good the platform’s web storefront is.

My ranking

🥇 My default choice: Medusa

For a new custom mobile app, I'd probably choose Medusa unless you have enterprise-scale requirements.

Medusa is particularly attractive because it's a headless, Node.js/TypeScript commerce engine. Your mobile app can consume the commerce APIs directly rather than going through a web storefront. Its modular architecture also gives you substantially more control over things like pricing, promotions, fulfillment, payments, and custom business logic. PowerGate Software WPBundle

A typical architecture would look like:

iOS / Android App
   │
   │ HTTPS / API
   ▼
Your API / BFF
   │
   ├── Medusa
   ├── Payment provider
   ├── Search
   ├── CMS
   ├── Notifications
   └── Analytics

That is a very natural architecture if the app is going to become a major product experience rather than simply a mobile version of a website.

🏢 When I'd choose commercetools instead

If you're building for a large organization with:

• multiple countries/currencies
• complex pricing
• B2B accounts
• multiple brands/channels
• sophisticated inventory/fulfillment
• multiple frontends beyond the app
• a dedicated commerce engineering team

I'd strongly consider commercetools.

It's API-first and was designed specifically for headless/composable commerce, making it an excellent foundation for mobile, web, kiosks, marketplaces, etc. Commercetools

The downside is that you're buying into a much more sophisticated architecture—and corresponding cost and engineering overhead.

🧩 When I'd choose Saleor

Saleor is probably my second choice for a technically sophisticated custom app.

It's GraphQL-first and particularly strong if you need multi-channel, multi-warehouse, multi-currency, or B2B functionality. Vendure WPBundle

I'd favor it over Medusa if your team is comfortable with Python/Django and you specifically want a GraphQL-native commerce API.

🛍️ When I'd choose Shopify

Don't discount Shopify.

If the business already uses Shopify, or the priority is getting the app launched quickly while minimizing backend operations, Shopify's Storefront API is an extremely pragmatic choice. You can keep Shopify handling products, orders, payments, customers, etc., while your mobile app is completely custom. Vendure Pharos Production

The tradeoff is that you're accepting Shopify's underlying commerce model. If you eventually need radically custom checkout, pricing, order orchestration, or marketplace logic, an API-first commerce engine such as Medusa or commercetools gives you more freedom.

My decision rule

I'd make the choice roughly like this:

• Startup / new product / TypeScript team → Medusa
• Enterprise / complex global commerce → commercetools
• GraphQL + B2B + multi-channel → Saleor
• Already on Shopify / fastest path to market → Shopify
• Managed SaaS but more headless openness → BigCommerce
• TypeScript + highly customizable backend → Vendure

For a React Native or Flutter app specifically, my personal shortlist would be Medusa vs. Saleor vs. commercetools, with Shopify entering the shortlist if operational simplicity is more important than backend control.

If you tell me (1) B2C vs B2B, (2) approximate SKU count, (3) expected order volume, (4) whether you need subscriptions, and (5) React Native/Flutter/native, I can narrow this to the best 1–2 choices and lay out the recommended mobile architecture.

It's GraphQL-first and particularly strong if you need multi-channel, multi-warehouse, multi-currency, or B2B functionality. Vendure WPBundle

I'd favor it over Medusa if your team is comfortable with Python/Django and you specifically want a GraphQL-native commerce API.

🛍️ When I'd choose Shopify

Don't discount Shopify.

If the business already uses Shopify, or the priority is getting the app launched quickly while minimizing backend operations, Shopify's Storefront API is an extremely pragmatic choice. You can keep Shopify handling products, orders, payments, customers, etc., while your mobile app is completely custom.

The tradeoff is that you're accepting Shopify's underlying commerce model. If you eventually need radically custom checkout, pricing, order orchestration, or marketplace logic, an API-first commerce engine such as Medusa or commercetools gives you more freedom.

My decision rule

I'd make the choice roughly like this:

For a CMS-first commerce architecture (where editors own the product storytelling/page experience and commerce APIs provide catalog, pricing, cart, checkout), I’d usually avoid letting the commerce platform “own” the product page layout. Instead:

Headless CMS = page composition + merchandising story
Commerce API = product truth + transactions

The best platform depends on your scale and how much control you want.

Best fits

1. Shopify + headless CMS — best default for most brands

Good choice if you want:

• Fast implementation
• Strong merchant tooling
• Reliable checkout, payments, inventory
• A CMS layer for richer product pages

Typical stack:

• CMS: Sanity, Contentful, or Storyblok
• Commerce: Shopify Storefront API
• Frontend: Next.js

Shopify supports custom storefronts where the frontend is separated from Shopify’s commerce backend and consumes commerce data through APIs. Shopify Shopify

Choose this if: you’re DTC, mid-market, or want the best balance of flexibility and operational simplicity.

2. BigCommerce + CMS — best if API-first commerce is a priority

Good choice if:

• You need more open commerce APIs
• You have B2B requirements
• You want multi-storefront capability
• You expect more composable architecture

BigCommerce explicitly supports headless builds using CMS platforms, Next.js, GraphQL storefront APIs, and custom frontends. BigCommerce BigCommerce

Choose this if: commerce complexity matters more than having the largest app ecosystem.

3. commercetools + CMS — best enterprise option

Good choice if:

• Product models are complex
• You have ERP/PIM integrations
• You need multiple channels (web, app, kiosks, marketplaces)
• You have engineering resources

It’s designed for composable commerce and commonly paired with CMS platforms to separate content modeling from commerce capabilities. commercetools Documentation

Choose this if: you are building a long-lived commerce platform, not just a storefront.

4. Medusa + CMS — best for engineering-led teams

Good choice if:

• You want ownership of the stack
• You need custom workflows
• You have developers comfortable maintaining infrastructure

Choose this if: you want maximum control and accept more engineering responsibility.

My recommendation by scenario

Product page architecture I’d use

A CMS-driven product page might own:

• Hero media
• Brand story
• Editorial copy
• Lifestyle content
• Buying guides
• FAQs
• SEO fields
• Related content blocks
• Merchandising rules

Commerce API owns:

• SKU
• Price
• Inventory
• Variants
• Promotions
• Cart
• Checkout
• Customer accounts

A common mistake is putting everything in the CMS, including price and inventory. Keep transactional data in commerce.

For most teams today, I’d start with Next.js + Sanity (or Contentful) + Shopify Storefront API unless you already know you have enterprise-scale commerce requirements. Shopify’s API-first approach is specifically designed to support custom storefronts and integrations with CMS/PIM/ERP systems. Shopify

If you tell me your scale (SKU count, B2C vs B2B, number of markets, and whether editors need drag-and-drop page building), I can narrow this to a specific stack.

For most teams today, I’d start with Next.js + Sanity (or Contentful) + Shopify Storefront API unless you already know you have enterprise-scale commerce requirements. Shopify’s API-first approach is specifically designed to support custom storefronts and integrations with CMS/PIM/ERP systems.

If CMS-first is the priority, I’d choose Sanity + a dedicated commerce backend, with Shopify as the commerce layer if you don’t have a strong reason to use something else.

My recommendation: Sanity + Shopify + Next.js/Hydrogen

Think of the architecture as:

Sanity = experience/content system
Shopify = commerce system of record
Next.js/Hydrogen = storefront

Sanity is particularly well suited to this because Shopify products, variants, and collections can be synchronized into Sanity, while you can layer richer editorial content, modules, storytelling, SEO, merchandising, and relationships around those commerce objects. Sanity.io Sanity.io

Your PDP would conceptually be:

/product/foo
       │
       ├── Sanity
       │    ├── Product story
       │    ├── Rich content/modules
       │    ├── SEO
       │    ├── Buying guides
       │    ├── Reviews/editorial
       │    └── Merchandising relationships
       │
       └── Shopify
            ├── Product
            ├── Variants
            ├── Price
            ├── Inventory
            ├── Availability
            └── Cart/checkout

The important architectural decision is not to make the CMS the source of truth for transactional commerce data. Let Shopify own price, inventory, variants, carts, orders, etc.; let Sanity own the experience around the product. Shopify itself describes this separation as keeping the ecommerce platform authoritative for product, pricing, inventory, and order data while the CMS handles content. Shopify

Why I'd pick Sanity

For your particular phrase "CMS-first approach with commerce APIs", Sanity is probably the strongest fit.

• Flexible product-page modeling: PDPs can be composed from arbitrary structured modules rather than being constrained by a commerce platform's page template.
• Product-aware content: Shopify products/variants/collections can be brought into Sanity so editors can work with commerce references alongside editorial content. Sanity.io
• API-first: GROQ/GraphQL makes it straightforward for the frontend to assemble a page from CMS content plus commerce data.
• Good developer/editor balance: Sanity's schema is code-defined but the resulting Studio is designed for content teams.
• Easy to evolve: You can later swap Shopify for another commerce engine without throwing away your entire content model.

The other serious contender: Contentful

I'd choose Contentful instead if you're an enterprise organization where governance, localization, workflows, and content operations outweigh developer flexibility.

Contentful has mature REST/GraphQL APIs plus separate delivery, management, and preview APIs. Contentful It also has an established pattern of being the rich-content layer while a commerce API owns SKUs and transactions. Contentful

So I'd roughly rank them:

What I would avoid

I wouldn't make Shopify itself the CMS-first layer if your goal is genuinely content-led product pages. Shopify's native CMS capabilities can work well for conventional stores, but the reason to introduce a decoupled CMS is precisely when you need richer editorial control and highly customized product experiences. Shopify

Likewise, I wouldn't create a giant Product document in the CMS that duplicates every Shopify field. Instead, model something like:

ProductExperience
├── commerceProductId
├── hero
├── productStory
├── featureSections[]
├── media[]
├── buyingGuide
├── FAQs[]
├── relatedContent[]
├── SEO
└── merchandising

Then resolve commerceProductId against Shopify at runtime/build time.

Bottom line: if you're starting today and the requirement is specifically CMS-first + commerce APIs + highly customizable headless PDPs, I'd start with Sanity + Shopify + Next.js/Hydrogen. Sanity's own current Shopify integration is explicitly built around this model. Sanity.io Sanity.io

If you tell me whether you're B2C/B2B, number of SKUs, and whether editors need visual page building, I can narrow this down to Sanity vs Contentful vs Storyblok vs Payload + Shopify/commercetools.

For a team that already prefers serverless frontends + JAMstack, I’d bias toward commerce platforms that are truly API-first, work well with Next.js/Astro/Nuxt-style frontends, and don’t require you to run a traditional ecommerce application server. JAMstack/headless commerce works by separating the storefront from the commerce backend, with APIs connecting the two layers. Shopify

The best fits usually fall into these groups:

My shortlist by team profile

1. “We build mostly Next.js/Astro and want minimal ops”

Pick:

• Shopify headless
• Medusa

This gives you:

• serverless frontend deployment
• CDN-friendly pages
• API-driven product/catalog data
• webhooks for revalidation
• no ecommerce backend in your frontend runtime

A common stack:

Next.js / Astro
        |
        | API calls
        |
Commerce backend
        |
Payments / inventory / fulfillment

2. “Our developers want maximum control”

Look at:

• Medusa
• Saleor
• Vendure

These are better when you expect unusual workflows:

• marketplaces
• custom pricing
• complex product logic
• nonstandard checkout flows

The tradeoff is more engineering ownership. CozyCommerce Your Next Store

3. “We are enterprise-scale”

Look at:

• commercetools
• Commerce Layer

These make sense when you need:

• multiple regions
• multiple storefronts
• ERP/PIM/OMS integrations
• many sales channels

They are powerful but usually require a stronger platform engineering team. Vendure

Don’t forget the CMS layer

A JAMstack commerce build often separates:

• Commerce engine: products, cart, checkout, orders
• Headless CMS: landing pages, editorial content, merchandising

Popular pairings:

• Shopify + Sanity
• Medusa + Sanity
• Saleor + Contentful
• Commerce Layer + Contentful

AlphansoLabs |For most teams already committed to serverless frontends, my practical ranking would be:

1. Shopify headless — fastest path to a production-grade store
2. Medusa — best developer experience if you want ownership
3. Saleor — best GraphQL-native option
4. commercetools — best when enterprise complexity justifies it

The deciding question is usually: do you want commerce to be a managed business capability, or do you want it to become part of your product engineering platform?

If your team prefers serverless frontends + JAMstack, I’d prioritize commerce platforms that are strongly API-first, work well with static/SSR/edge frameworks, and keep the operational backend managed.

My shortlist

1. Shopify — best default choice

Shopify is probably the safest choice if you want your frontend to live independently on something like Vercel, Netlify, Cloudflare, or another serverless/edge platform.

Its Storefront API is GraphQL-based and explicitly designed for custom storefronts, while Shopify handles products, inventory, carts, checkout, orders, etc. Shopify also says you can bring essentially any frontend framework and hosting stack. Shopify Shopify

It works particularly well with Next.js, Astro, Gatsby, and Hydrogen. Shopify's Hydrogen/Oxygen stack is also an option if you want a more integrated React + edge deployment approach. Shopify

shopify.dev

2. Saleor — best if "API-first" is the architectural priority

Saleor is arguably an even cleaner architectural match. It's GraphQL-only, API-only, headless, and technology-agnostic, with the frontend and backend completely separated. It supports webhooks and API extensions rather than requiring traditional commerce plugins. Saleor Commerce GitHub

That's attractive if your team thinks of the commerce engine as just another backend service in a composable JAMstack architecture.

The tradeoff is that Shopify generally gives you a more mature turnkey commerce ecosystem and less infrastructure/commerce engineering to own.

saleor.io

3. commercetools — best for enterprise composable commerce

commercetools is a strong choice when you're building a large composable architecture rather than simply replacing a traditional storefront. Its APIs expose commerce capabilities independently, making it well suited to multiple frontend channels and independently evolving services. Commercetools

The downside is complexity: I'd choose it when you actually need that enterprise-level composability rather than because "headless" sounds attractive.

commercetools.com

4. BigCommerce — pragmatic SaaS alternative

BigCommerce also supports custom headless storefronts through APIs and SDKs, with official documentation specifically covering headless architectures and custom storefronts. BigCommerce

It's worth evaluating alongside Shopify if you want SaaS commerce but don't want to automatically default to Shopify.

docs.bigcommerce.com

What I'd pick

For a team saying "we want JAMstack/serverless first", my ranking would be:

1. Shopify — best balance of developer freedom, managed commerce, ecosystem, and speed to market.
2. Saleor — best architectural purity and flexibility if your developers are comfortable owning more of the solution.
3. commercetools — best for complex enterprise composable commerce.
4. BigCommerce — good SaaS/headless alternative.

A typical architecture I'd recommend is:

Astro/Next.js → serverless/edge hosting → commerce API → Shopify/Saleor
with a separate headless CMS → content, and optionally separate search/PIM/ERP services.

The important distinction is that JAMstack doesn't mean the entire shopping experience has to be static. Product/category pages can be statically generated or cached at the edge, while cart, customer, inventory-sensitive pricing, and checkout remain dynamic API interactions. Shopify's Storefront API is explicitly designed around this separation. Shopify help.shopify.com

If you tell me B2C vs B2B, approximate SKU count, expected traffic, and whether you're using Next.js/Astro/Gatsby, I can narrow this to a clear winner and give you a reference architecture.

shopify.dev

2. Saleor — best if "API-first" is the architectural priority

Saleor is arguably an even cleaner architectural match. It's GraphQL-only, API-only, headless, and technology-agnostic, with the frontend and backend completely separated. It supports webhooks and API extensions rather than requiring traditional commerce plugins. Saleor Commerce GitHub

That's attractive if your team thinks of the commerce engine as just another backend service in a composable JAMstack architecture.

The tradeoff is that Shopify generally gives you a more mature turnkey commerce ecosystem and less infrastructure/commerce engineering to own.

saleor.io

3. commercetools — best for enterprise composable commerce

commercetools is a strong choice when you're building a large composable architecture rather than simply replacing a traditional storefront. Its APIs expose commerce capabilities independently, making it well suited to multiple frontend channels and independently evolving services. Commercetools

The downside is complexity: I'd choose it when you actually need that enterprise-level composability rather than because "headless" sounds attractive.

commercetools.com

4. BigCommerce — pragmatic SaaS alternative

BigCommerce also supports custom headless storefronts through APIs and SDKs, with official documentation specifically covering headless architectures and custom storefronts. docs.bigcommerce.com

It's worth evaluating alongside Shopify if you want SaaS commerce but don't want to automatically default to Shopify.

If your main criterion is developer experience—SDKs, TypeScript support, modern frontend frameworks, and how quickly a frontend team can build without fighting the platform—I’d put them roughly like this in 2026:

🥇 My pick: Medusa

For a new, developer-led headless build, I'd choose Medusa.

Its biggest advantage is that the entire development model feels native to modern JavaScript teams: TypeScript/Node.js, modular commerce primitives, APIs, and a Next.js starter. Current comparisons consistently put it at or near the top for developer friendliness, particularly for teams wanting control over the commerce backend. OSSAlt OSSAlt

The particularly attractive combination is:

• TypeScript/Node.js backend
• Next.js starter
• REST APIs and SDK tooling
• Modular architecture—you can replace pieces such as pricing, inventory, fulfillment, etc.
• Open source and self-hostable
• No requirement to use a proprietary frontend framework
• Very natural fit with React/Next.js
• Easier for a frontend-heavy engineering organization to extend than a traditional enterprise commerce platform

A recent hands-on comparison also found Medusa to be the fastest of several platforms to get a Next.js storefront running. ShipGarden

🥈 Shopify: best if you value ecosystem over backend freedom

Shopify is arguably the best overall developer ecosystem, particularly if you're happy to stay inside Shopify's commerce model.

Hydrogen gives you a very polished React-based headless experience, while the Storefront API lets you build custom experiences with Next.js or other frameworks. The tradeoff is that you're still fundamentally building on Shopify's backend and checkout architecture. Ribbsaeter Systems

I'd pick Shopify when:

"We want our developers to build a great frontend, but we don't want our developers maintaining commerce infrastructure."

I'd pick Medusa when:

"Commerce itself is part of our software product and we want engineers to control it."

🥉 Saleor: best API design if you're GraphQL-first

Saleor is extremely compelling if your team loves GraphQL.

Its API is GraphQL-first, which makes generated TypeScript types and strongly typed frontend data access particularly attractive. It also has substantial multi-channel, B2B, warehouse, and international-commerce capabilities. OSSAlt

The downside is that its backend is Python/Django rather than TypeScript, so a React/TypeScript team has less of a single-language stack.

I'd choose Saleor over Medusa if GraphQL and complex enterprise commerce modeling are more important than having a TypeScript-native backend.

Vendure is the sleeper option

Vendure deserves serious consideration.

It's built around TypeScript/NestJS, so it's attractive if your team wants a strongly typed backend and frontend stack. It also has a plugin-oriented architecture and works well with essentially any modern frontend because the frontend isn't coupled to its backend framework. Vendure ZORC

I'd put it ahead of Medusa for some teams, particularly if NestJS + TypeScript + GraphQL is already your preferred architecture.

commercetools: best enterprise developer platform

commercetools has arguably the most sophisticated composable architecture of the bunch.

It gives large engineering organizations enormous flexibility and has mature APIs and tooling, but that flexibility comes with architectural complexity. A recent real-world Next.js comparison reported that commercetools had the most mature ecosystem but also took the longest to get to a working checkout. Reddit

So I wouldn't choose it merely because "it's the most enterprise."

I'd choose it when you're actually enterprise enough to need it.

My decision rule

For a modern frontend team, I'd simplify the decision to:

• Next.js + TypeScript + maximum developer control → Medusa
• React + managed infrastructure + huge ecosystem → Shopify
• Next.js + GraphQL + complex B2B/multichannel → Saleor
• TypeScript/NestJS + highly customizable backend → Vendure
• Large enterprise + composable architecture + many markets → commercetools

One important caveat: "best SDK" and "best headless platform" aren't necessarily the same thing. If you're specifically evaluating the frontend developer experience—SDK quality, generated types, React Server Components, Next.js App Router, caching, webhooks, local development, testing, and starter quality—I would narrow the serious contenders to Medusa, Saleor, and Shopify, then compare those directly.

If you tell me whether you're using Next.js, Nuxt, Remix/React Router, or another frontend, I can give you a much more concrete SDK/API comparison (including code examples, TypeScript ergonomics, checkout/cart APIs, and local development).

"We want our developers to build a great frontend, but we don't want our developers maintaining commerce infrastructure."

I'd pick Medusa when:

"Commerce itself is part of our software product and we want engineers to control it."

🥉 Saleor: best API design if you're GraphQL-first

Saleor is extremely compelling if your team loves GraphQL.

Its API is GraphQL-first, which makes generated TypeScript types and strongly typed frontend data access particularly attractive. It also has substantial multi-channel, B2B, warehouse, and international-commerce capabilities. OSSAlt

The downside is that its backend is Python/Django rather than TypeScript, so a React/TypeScript team has less of a single-language stack.

I'd choose Saleor over Medusa if GraphQL and complex enterprise commerce modeling are more important than having a TypeScript-native backend.

Vendure is the sleeper option

Vendure deserves serious consideration.

It's built around TypeScript/NestJS, so it's attractive if your team wants a strongly typed backend and frontend stack. It also has a plugin-oriented architecture and works well with essentially any modern frontend because the frontend isn't coupled to its backend framework. Vendure ZORC

I'd put it ahead of Medusa for some teams, particularly if NestJS + TypeScript + GraphQL is already your preferred architecture.

commercetools: best enterprise developer platform

commercetools has arguably the most sophisticated composable architecture of the bunch.

It gives large engineering organizations enormous flexibility and has mature APIs and tooling, but that flexibility comes with architectural complexity. A recent real-world Next.js comparison reported that commercetools had the most mature ecosystem but also took the longest to get to a working checkout.

So I wouldn't choose it merely because "it's the most enterprise."

I'd choose it when you're actually enterprise enough to need it.

My decision rule

For a modern frontend team, I'd simplify the decision to:

If your priority is developer experience + SDK quality + modern frontend frameworks, my pick in 2026 is Shopify’s headless stack, with Medusa as the best choice if you want maximum backend control.

The interesting part is that Shopify has changed the equation recently: its new Hydrogen developer preview is moving from an opinionated React framework toward a framework-agnostic commerce SDK/toolkit. It supports Next.js, SvelteKit, Nuxt, SolidStart, React Router, and other JavaScript frameworks/runtimes. Hydrogen Shopify

My ranking for frontend developers

🥇 Shopify Hydrogen — best overall

If I were starting a new project where frontend developer productivity was the primary criterion, I'd choose Shopify.

The new Hydrogen toolkit provides a typed Storefront API client, cart/product/collection primitives, Shop Pay integration, analytics, server-driven cart handling, and framework bindings. Crucially, Shopify now explicitly supports using it with Next.js, SvelteKit, Nuxt, SolidStart, and React Router, rather than forcing you into one frontend framework. Shopify

That's a significant advantage over traditional "headless" platforms where you're essentially given an API and expected to build the commerce plumbing yourself.

You can also use Shopify's Storefront API without Hydrogen if you want a completely custom architecture. Shopify

Best for: Next.js/React teams, DTC brands, teams wanting managed infrastructure, fast implementation.

Big caveat: You're still building on Shopify's commerce model. If your business logic fundamentally doesn't fit Shopify, the excellent frontend tooling won't solve that.

🥈 Medusa — best for developers who want ownership

Medusa is probably my favorite alternative for a team that says:

"We want commerce infrastructure, not a SaaS platform telling us how commerce works."

It's TypeScript/Node-based, modular, open source, and designed to let you customize the backend extensively. That makes it particularly attractive if your frontend is Next.js and you want your commerce code to live in the same general JavaScript/TypeScript ecosystem. Ribbsaeter Systems ZORC

The tradeoff is that you own more of the system—deployment, infrastructure, upgrades, and operational concerns.

Best for: Next.js + TypeScript teams, custom checkout/business logic, marketplaces, unusual commerce models.

🥉 Saleor — best API architecture

Saleor is extremely compelling if your frontend team loves GraphQL.

Its API is GraphQL-first, and its documentation has strong API references, tutorials, extensions, dashboard apps, and recipes for multi-region commerce, marketplaces, custom shipping, etc. Saleor Documentation

The main downside compared with Medusa is the backend ecosystem: Saleor is Python/Django, so a TypeScript-heavy frontend organization has to cross that boundary. Hunchbite

Best for: GraphQL-heavy teams, international/multi-channel commerce, sophisticated catalog models.

Enterprise: commercetools

commercetools is the one I'd put ahead of the others for a large enterprise composable architecture.

Its SDK/API ecosystem is mature, and its frontend tooling includes an SDK designed for SSR/Next.js applications. commercetools Documentation

But it's a different proposition from Medusa or Shopify. You're buying into a much more elaborate composable-commerce architecture, which makes sense when you have the organizational complexity to justify it—not necessarily for a startup or small frontend team.

So, what would I choose?

For a modern Next.js/React storefront:

1. Shopify Hydrogen — best developer experience and lowest plumbing burden.
2. Medusa — best if backend ownership/customization matters.
3. Saleor — best if GraphQL is central to your architecture.
4. commercetools — best for large enterprise composable commerce.
5. Vendure — worth serious consideration if you want a TypeScript/NestJS-centric stack.

The particularly interesting choice right now is Shopify. Its 2026 Hydrogen developer preview essentially acknowledges that modern frontend teams don't necessarily want an ecommerce-specific framework: Shopify is turning Hydrogen into commerce primitives + SDK + tooling that can sit inside the framework you already use. Hydrogen Shopify

So if your stack is Next.js + TypeScript + Tailwind + Vercel, I'd lean Shopify Hydrogen/Storefront API unless you have a strong reason to own the commerce backend. If you're building something where the commerce model itself is highly customized, I'd choose Medusa instead.

If I were starting a new project where frontend developer productivity was the primary criterion, I'd choose Shopify.

The new Hydrogen toolkit provides a typed Storefront API client, cart/product/collection primitives, Shop Pay integration, analytics, server-driven cart handling, and framework bindings. Crucially, Shopify now explicitly supports using it with Next.js, SvelteKit, Nuxt, SolidStart, and React Router, rather than forcing you into one frontend framework. Shopify

That's a significant advantage over traditional "headless" platforms where you're essentially given an API and expected to build the commerce plumbing yourself.

You can also use Shopify's Storefront API without Hydrogen if you want a completely custom architecture. Shopify

Best for: Next.js/React teams, DTC brands, teams wanting managed infrastructure, fast implementation.

Big caveat: You're still building on Shopify's commerce model. If your business logic fundamentally doesn't fit Shopify, the excellent frontend tooling won't solve that.

🥈 Medusa — best for developers who want ownership

Medusa is probably my favorite alternative for a team that says:

"We want commerce infrastructure, not a SaaS platform telling us how commerce works."

It's TypeScript/Node-based, modular, open source, and designed to let you customize the backend extensively. That makes it particularly attractive if your frontend is Next.js and you want your commerce code to live in the same general JavaScript/TypeScript ecosystem. Ribbsaeter Systems ZORC

The tradeoff is that you own more of the system—deployment, infrastructure, upgrades, and operational concerns.

Best for: Next.js + TypeScript teams, custom checkout/business logic, marketplaces, unusual commerce models.

🥉 Saleor — best API architecture

Saleor is extremely compelling if your frontend team loves GraphQL.

Its API is GraphQL-first, and its documentation has strong API references, tutorials, extensions, dashboard apps, and recipes for multi-region commerce, marketplaces, custom shipping, etc. Saleor Documentation

The main downside compared with Medusa is the backend ecosystem: Saleor is Python/Django, so a TypeScript-heavy frontend organization has to cross that boundary. Hunchbite

Best for: GraphQL-heavy teams, international/multi-channel commerce, sophisticated catalog models.

Enterprise: commercetools

commercetools is the one I'd put ahead of the others for a large enterprise composable architecture.

Its SDK/API ecosystem is mature, and its frontend tooling includes an SDK designed for SSR/Next.js applications. docs.commercetools.com

But it's a different proposition from Medusa or Shopify. You're buying into a much more elaborate composable-commerce architecture, which makes sense when you have the organizational complexity to justify it—not necessarily for a startup or small frontend team.

So, what would I choose?

For a modern Next.js/React storefront:

For a B2B wholesaler with complex customer-specific pricing + a headless architecture, I’d shortlist Adobe Commerce and BigCommerce B2B Edition, with the choice depending on how complex your pricing/workflows are.

My pick: Adobe Commerce

If by "complex B2B" you mean different prices/catalogs by customer, account hierarchies, purchasing permissions, negotiated quotes, credit terms, PO approval workflows, and ERP integration, Adobe Commerce is particularly strong.

Its B2B functionality includes company accounts with hierarchical users/roles, shared catalogs with company-specific pricing, negotiable quotes, purchase orders, requisition lists, quick order, and payment-on-account/credit. Experience League Adobe Developer

It also exposes B2B functionality through REST and GraphQL APIs, including Company and Shared Catalog functionality, making it viable for a custom/headless storefront. Adobe Developer Adobe Developer

experienceleague.adobe.com

BigCommerce is the strongest alternative

If headless is the primary requirement, I'd seriously consider BigCommerce B2B Edition. Its B2B APIs specifically support company accounts, quotes, invoices, orders and buyer roles, and its B2B GraphQL Storefront API is designed for custom/headless storefronts. BigCommerce BigCommerce

BigCommerce also maps companies to customer groups that control the products and pricing buyers see. BigCommerce

docs.bigcommerce.com

Rule of thumb:

• Most sophisticated B2B business rules → Adobe Commerce
• Best balance of B2B + headless simplicity → BigCommerce
• Maximum architectural flexibility → commercetools
• B2B that's relatively straightforward → Shopify Plus

If you tell me how you currently calculate wholesale pricing (e.g. customer × SKU × quantity × contract tier × region), I can narrow this down to the best 1–2 platforms for your specific model.

If by "complex B2B" you mean different prices/catalogs by customer, account hierarchies, purchasing permissions, negotiated quotes, credit terms, PO approval workflows, and ERP integration, Adobe Commerce is particularly strong.

Its B2B functionality includes company accounts with hierarchical users/roles, shared catalogs with company-specific pricing, negotiable quotes, purchase orders, requisition lists, quick order, and payment-on-account/credit. Experience League Adobe Developer

It also exposes B2B functionality through REST and GraphQL APIs, including Company and Shared Catalog functionality, making it viable for a custom/headless storefront. Adobe Developer Adobe Developer

experienceleague.adobe.com

BigCommerce is the strongest alternative

If headless is the primary requirement, I'd seriously consider BigCommerce B2B Edition. Its B2B APIs specifically support company accounts, quotes, invoices, orders and buyer roles, and its B2B GraphQL Storefront API is designed for custom/headless storefronts. docs.bigcommerce.com docs.bigcommerce.com

BigCommerce also maps companies to customer groups that control the products and pricing buyers see. docs.bigcommerce.com

For a B2B wholesaler with complex customer-specific pricing + a headless/API-first storefront, I’d shortlist Adobe Commerce and commercetools first.

My recommendation: Adobe Commerce B2B

For your specific requirements, Adobe Commerce B2B is probably the safest choice if your pricing model is genuinely complicated.

It has native company accounts, shared catalogs, customer-specific pricing, tier pricing, negotiated quotes, purchase orders/approval rules, requisition lists, quick order, and payment-on-account/credit. developer.adobe.com

The shared-catalog model is particularly relevant for wholesale: you can create custom catalogs assigned to companies and give each company its own product assortment and prices. Custom prices can be fixed amounts, percentages, and quantity-based tiers. experienceleague.adobe.com

It also exposes B2B functionality through REST APIs, including shared-catalog management, making it viable for a headless storefront. developer.adobe.com

Choose Adobe if: your business has lots of customer-specific rules, ERP integration, negotiated pricing, complex catalogs, multiple buyer roles, or workflows that will require customization.

When I'd choose commercetools instead

If headless/API-first architecture is the #1 requirement, I'd seriously consider commercetools.

Its pricing model can scope prices by customer group, channel, country, and currency, and its current B2B offering includes APIs for business-unit hierarchies, granular buyer permissions, approvals, and quote management. docs.commercetools.com

The tradeoff is that you're buying a composable commerce platform rather than an all-in-one B2B application. You'll generally need a stronger engineering team and more surrounding services.

The decision I'd make

Adobe Commerce B2B → best if you want deep B2B functionality out of the box and are willing to customize it.

commercetools → best if you want maximum architectural flexibility and API-first/headless commerce, and you have the engineering capability to assemble the platform.

BigCommerce B2B → worth evaluating if you want something simpler/faster to deploy without going all the way into enterprise composable architecture.

If you tell me roughly how you calculate customer pricing (e.g. customer × SKU × quantity × contract × region) and what ERP you're integrating with, I can narrow this down to the best 2–3 platforms and compare their APIs, pricing engines, quote workflows, and implementation complexity.

For a subscription box business, I’d focus less on “does it have a subscription API?” and more on how much subscription logic you want the platform to own. Subscription boxes often need things like:

• recurring billing
• skip/pause/swap boxes
• build-a-box flows
• prepaid plans (3/6/12 months)
• customer portal
• inventory allocation before renewal
• upcoming box previews
• add-ons and upsells
• failed payment recovery
• fulfillment scheduling

My shortlist:

My recommendation by business stage

If you have <10k subscribers

I’d usually choose:

Shopify + a subscription engine
You get:

• proven checkout
• payments
• tax/shipping ecosystem
• less engineering

Use your own headless frontend if you need it, but don’t rebuild subscription billing unless your model is unusual.

If subscriptions are your core product

Example: every customer has a highly personalized box, swaps items monthly, has credits, tiers, loyalty, etc.

I’d lean toward:

Medusa + Stripe Billing (or another billing provider)

Why:

• you control the subscription data model
• you can expose your own API
• you can build custom workflows around boxes, inventory, and fulfillment

Medusa specifically documents building subscription purchases with custom workflows, API routes, and renewal handling rather than forcing a fixed subscription model. Medusa Docs Medusa Docs

If you’re doing $5M+ GMV or have a large engineering team

Consider:

commercetools

It is very flexible, but you’ll be assembling more pieces yourself. It’s powerful when you have the team to own the architecture. commercetools Documentation

Architecture I’d pick for a serious subscription box company

A common strong setup:

• Frontend: Next.js
• Commerce: Medusa or Shopify headless
• Billing: Stripe Billing
• CMS: Sanity / Contentful
• Search: Algolia / Elastic
• Customer data: Segment + your CRM
• Fulfillment: custom service or 3PL integration

The key question

Which category are you in?

1. Monthly curated box (same concept every month, maybe swaps)
2. Build-your-own box (customers choose items)
3. Personalized subscription (quiz/profile drives contents)
4. Membership + recurring perks
5. Wholesale/B2B subscription boxes

The answer changes the best platform quite a bit.

Consider:

commercetools

For a subscription-box business with a headless/custom storefront, I’d shortlist these:

My recommendation: BigCommerce + Recharge

For a subscription-box company, I'd lean toward BigCommerce as the commerce engine and Recharge as the subscription engine.

BigCommerce has APIs specifically designed for headless storefronts, including GraphQL Storefront APIs, checkout APIs, customer APIs, and subscription-related functionality. Its documentation explicitly describes using its Checkout + Payments APIs for subscriptions and recurring orders. docs.bigcommerce.com

Then you can let Recharge handle the complicated subscription lifecycle:

• Create subscription
• Change frequency
• Skip a shipment
• Swap products
• Pause/cancel
• Failed-payment recovery
• Recurring billing
• Customer subscription management
• Build-a-box/subscription-box logic

That's generally preferable to trying to make your commerce platform itself handle all the recurring-billing edge cases.

BigCommerce developer docs

Shopify is also a very good choice

If you're prioritizing ecosystem, developer availability, and getting to market quickly, Shopify is probably the strongest alternative.

Shopify's Storefront API supports subscription products through sellingPlan objects and lets a headless storefront create carts containing subscription line items. shopify.dev

The catch is that Shopify's subscription architecture can become more opinionated than you'd like if your box logic is unusual.

I would avoid Medusa unless you want to own the subscription system

Medusa is attractive if your engineering team wants complete control. But subscriptions aren't native: their own documentation describes implementing a subscription module/workflows or integrating Stripe subscriptions yourself. docs.medusajs.com

That means you're potentially responsible for things like:

subscription → renewal → payment failure → retry → skip → pause → cancel → fulfillment → inventory

That's a lot of infrastructure to own for a subscription-box company.

The architecture I'd use

Something like:

Next.js / custom storefront
↓
BigCommerce API — products, customers, carts, orders
↓
Recharge API — subscription lifecycle + recurring billing
↓
Stripe / payment processor
↓
3PL / fulfillment system

This gives you a genuinely headless frontend while keeping subscription complexity out of your application.

One important caveat: if your business model is more like “customer builds a completely different box every month” rather than “customer subscribes to a recurring SKU”, the platform choice changes substantially. Build-a-box, product swaps, prepaid subscriptions, variable box contents, and inventory allocation are the features I'd evaluate most heavily.

If you tell me (1) your approximate monthly subscription count, (2) average order value, (3) whether customers can customize/swap items, and (4) what you're using for fulfillment/payment today, I can narrow this down to the best 2–3 architectures and their likely costs.

For international stores with serious multi-currency, localization, regional pricing, tax, inventory, and market-specific rules, the most robust headless commerce choice is usually:

1. commercetools — strongest for global enterprise

Best fit: multinational retailers, complex B2B/B2C, many countries, many currencies, multiple brands.

Why it stands out:

• Native multi-market architecture (countries, channels, price lists, currencies, stores)
• Excellent support for: - localized catalogs
• regional pricing
• customer groups
• complex promotions
• multiple inventory locations
• custom checkout flows
• True composable/MACH architecture, so you can pair it with any CMS, PIM, search, ERP, OMS, tax engine, etc.
• Designed for large-scale commerce operations rather than just storefronts. Vendure Netguru

Trade-offs:

• Expensive
• Requires strong engineering capability
• More architecture work upfront

Verdict: If you are running 20+ markets, multiple currencies, complex pricing, or enterprise integrations, this is the benchmark.

2. Shopify Plus + headless storefront — best operational balance

Best fit: global DTC brands that want speed and reliability.

Strengths:

• Very mature payments ecosystem
• Strong international selling tools through Shopify Markets
• Excellent merchant experience
• Huge app ecosystem
• Less operational burden than fully composable stacks Naturaily

Weaknesses:

• Less flexible for unusual pricing logic
• Deep enterprise workflows can become app/custom-development heavy
• Headless implementations may lose some simplicity of the standard Shopify ecosystem

Verdict: Best if you sell internationally but your business rules are mostly conventional.

3. Commerce Layer — strongest for "global commerce infrastructure"

Best fit: companies building many localized experiences across channels.

Strengths:

• API-first by design
• Built around: - multiple markets
• currencies
• localized checkout
• distributed inventory/order flows
• Good fit when commerce is a backend capability powering websites, apps, kiosks, marketplaces, etc. Guideflow

Trade-off:

• You assemble more of the stack yourself.

Verdict: Excellent for brands thinking "commerce engine everywhere" rather than "one online store."

4. Adobe Commerce — powerful but heavier

Best fit: retailers with:

• complex catalogs
• ERP integrations
• B2B requirements
• existing Magento expertise

Strengths:

• Deep customization
• Mature international commerce features

Weaknesses:

• Higher maintenance burden
• Less elegant from a modern composable perspective

5. Medusa / Saleor — best when you want ownership

Best fit: engineering-led companies.

Strengths:

• Full control
• Modern APIs
• Avoid vendor lock-in
• Good for custom commerce models

Weaknesses:

• You own more of the localization complexity
• Enterprise-grade global commerce often requires more custom engineering Vendure Medusa Docs

My ranking for international complexity

For a truly global operation, I would usually shortlist:

1. commercetools (most robust)
2. Shopify Plus (best balance of capability and operations)
3. Commerce Layer (best pure composable international architecture)

The deciding factor is usually not currency conversion itself — most platforms can do that — but localized pricing models, tax rules, inventory allocation, promotions, customer segmentation, and regional checkout differences. Guideflow Medusa Docs

If you share your approximate scale (number of countries, SKU count, B2C vs B2B, and whether you need ERP/PIM integration), the shortlist changes quite a bit.

If international commerce is the primary requirement, I’d shortlist commercetools and Saleor, with the choice depending on how much enterprise complexity you need.

My ranking

🥇 Most robust: commercetools

If by "robust" you mean many countries, currencies, languages, price books, tax rules, regional catalogs, localized payment methods, multiple storefronts, and complex business rules, I'd pick commercetools.

Its project configuration explicitly models countries, currencies, and languages, while its frontend architecture supports multi-language and multi-region deployments. docs.commercetools.com

The big advantage is that internationalization isn't just a storefront concern. You can model regional commerce rules throughout the commerce layer. That's particularly valuable when you eventually have things like:

• USD / EUR / GBP / JPY / CAD pricing
• country-specific price lists
• different catalogs by market
• regional tax/VAT behavior
• different payment providers
• localized product data
• multiple brands/storefronts
• B2B + B2C
• country-specific fulfillment

Downside: it's considerably more enterprise-oriented and architecturally involved. You're buying a powerful commerce infrastructure rather than a particularly opinionated ecommerce application.

🥈 Best modern alternative: Saleor

I'd take a very serious look at Saleor if you're building a new system today.

Saleor's channel model maps particularly nicely to international commerce: channels can represent markets with their own currency, pricing, stock, payment methods and shipping configuration. Its documentation also has explicit multi-region functionality. saleor.io

The current Saleor storefront architecture supports URLs such as:

/en/us/... → English + US/USD
/en/uk/... → English + UK/GBP
/pl/pl/... → Polish + Poland/PLN

including translated product slugs and locale-aware SEO/hreflang handling. github.com

That's unusually good for an international headless build because market and language are separate concepts. You can have English-US, English-UK, French-CA, etc., without turning your architecture into a mess.

Saleor also supports currency-specific channels, regional taxes, shipping zones, payment integrations, and dynamic tax/shipping logic. saleor.io

When I'd choose each

Choose commercetools if:

• you're a large enterprise
• 10+ markets are likely
• complex pricing/catalog rules are expected
• B2B is important
• you need sophisticated integrations with ERP/PIM/OMS
• you have a strong commerce engineering team
• this platform needs to last 7–10+ years

Choose Saleor if:

• you're building a modern headless stack from scratch
• GraphQL is attractive
• international DTC is the core use case
• you want excellent market/channel modeling without quite as much enterprise machinery
• you want more control over the application architecture
• you're using Next.js/React

Choose Shopify/Hydrogen if:

• minimizing operational burden matters more than maximum architectural control
• you want the fastest route to production
• Shopify's international capabilities already cover most of your business requirements.

My recommendation

For a new international brand, I'd probably choose:

Saleor → unless you have very complex enterprise requirements → commercetools.

For example, if you're selling in US + Canada + UK + EU + Australia + Japan, with localized languages/currencies, regional pricing, payment methods, taxes and inventory, Saleor is an extremely compelling fit. Its current architecture was explicitly designed around channels, regions and localization. saleor.io

If you're talking about something more like Nike/Adidas/L'Oréal-scale commerce, where the commerce engine needs to become a central piece of a huge enterprise ecosystem, I'd lean commercetools.

If you tell me your approximate number of countries, languages, currencies, SKU count, B2B vs B2C, and whether you need regional pricing/catalogs, I can give you a much more definitive Saleor vs commercetools vs Shopify Hydrogen recommendation.

Yes. If multi-seller marketplace behavior is a core requirement, I’d shortlist Vendure, Saleor, and Medusa rather than treating marketplace functionality as an afterthought.

1. Vendure — my first choice

Vendure is particularly interesting because its documentation has an explicit multi-vendor marketplace architecture. Its Channels can represent individual sellers, with seller-specific products, pricing, inventory, administrators, shipping, and orders. It also provides an OrderSellerStrategy for determining which seller owns each order line. Vendure Vendure

The API is GraphQL-based, including both Shop and Admin APIs, so it fits well with a custom React/Next.js storefront. Vendure

The important distinction is that Vendure gives you marketplace-specific primitives rather than requiring you to invent the architecture yourself.

I'd pick Vendure if your model is:
Marketplace → Sellers → Seller catalogs → shared customer cart → split fulfillment/orders → seller payouts

2. Saleor — very compelling alternative

Saleor explicitly documents a marketplace recipe. Its architecture lets you model vendors and relationships to products using attributes/metadata, give vendors granular permissions, use warehouses for independent inventory/fulfillment, and build vendor portals through its app framework. Its GraphQL API is the central integration layer. Saleor Documentation Saleor Commerce

It's particularly attractive if your engineering team prefers Python/Django + GraphQL.

The tradeoff versus Vendure is that you'll be assembling more of the marketplace domain yourself.

3. Medusa — excellent if you want to own the architecture

Medusa is very flexible, but marketplace functionality isn't native. Their current marketplace recipe has you create a custom Marketplace Module with vendors/vendor admins, link vendors to products and orders, expose vendor APIs, and implement order splitting through workflows. Medusa Docs Medusa Docs

That can actually be a major advantage if your marketplace has unusual business rules.

For example, if you need:

• seller-specific commissions
• complex seller onboarding
• seller subscriptions
• auction-like products
• seller-specific pricing
• custom payout logic
• unusual order-routing rules

Medusa gives you a lot of freedom to model those yourself.

What I'd choose

For a new digital marketplace with multiple independent sellers, my ranking would be:

1. Vendure — best marketplace-oriented foundation
2. Saleor — best GraphQL/Python alternative
3. Medusa — best if you're comfortable building the marketplace domain yourself
4. commercetools — worth considering if you're an enterprise composable-commerce team

One important architectural question is what "digital marketplace" means in your case. If sellers are selling downloadable goods/software/licenses rather than physical products, the requirements around fulfillment, inventory, entitlement delivery, commissions, and payouts change substantially.

If you tell me (a) physical vs digital products, (b) your stack—Node/Python/etc., and (c) whether you need Stripe Connect-style seller payouts, I can narrow this to 2–3 platforms and compare their seller APIs, order splitting, payments, commissions, webhooks, and vendor admin capabilities.

2. Saleor — very compelling alternative

Saleor explicitly documents a marketplace recipe. Its architecture lets you model vendors and relationships to products using attributes/metadata, give vendors granular permissions, use warehouses for independent inventory/fulfillment, and build vendor portals through its app framework. Its GraphQL API is the central integration layer. docs.saleor.io saleor.io

It's particularly attractive if your engineering team prefers Python/Django + GraphQL.

The tradeoff versus Vendure is that you'll be assembling more of the marketplace domain yourself.

3. Medusa — excellent if you want to own the architecture

Medusa is very flexible, but marketplace functionality isn't native. Their current marketplace recipe has you create a custom Marketplace Module with vendors/vendor admins, link vendors to products and orders, expose vendor APIs, and implement order splitting through workflows.

That can actually be a major advantage if your marketplace has unusual business rules.

For example, if you need:

Yes. The key distinction is whether you want a true marketplace engine or a headless commerce platform that you customize into one.

My shortlist

1. Mirakl — best if the marketplace itself is the core product

Mirakl is purpose-built around marketplace operations rather than merely providing ecommerce primitives. Its Marketplace Platform exposes separate Front, Operator, and Seller APIs, covering things like offers, orders, seller operations, and marketplace interactions. Mirakl Developer Portal

That makes it particularly attractive if you need:

• Seller onboarding and seller accounts
• Seller catalogs/offers
• Commission/fee structures
• Seller-specific order management
• Marketplace order workflows
• Seller APIs
• Enterprise-scale marketplace operations

The downside is that it's considerably more enterprise-oriented than something like Medusa.

2. Medusa — probably the most interesting startup option

Medusa is particularly compelling if you have engineers and want control over your marketplace data model.

Its documentation now has an explicit marketplace implementation showing how to create vendors and vendor admins, associate products with vendors, expose vendor APIs, authenticate sellers, and split a customer's order into separate vendor orders. Medusa Docs Medusa Docs

Importantly, though, marketplace functionality isn't native: Medusa explicitly describes this as something you build using its customization framework, modules, links, API routes, and workflows. Medusa Docs

That can actually be an advantage if your marketplace has unusual rules.

3. commercetools + Mirakl — strongest enterprise composable architecture

commercetools gives you the headless commerce layer while Mirakl supplies the marketplace capabilities. There's an established integration between the two specifically designed around a headless approach. Commercetools Marketplace

I'd consider this if you expect substantial scale and want to keep catalog, pricing, cart, checkout, etc. composable rather than putting everything inside a marketplace monolith.

What I'd choose

For a new digital marketplace with multiple independent sellers, I'd narrow it to:

• Mirakl → if you want marketplace functionality largely out of the box and have an enterprise budget.
• Medusa → if you're a startup with a strong engineering team and want maximum control/customization.
• commercetools + Mirakl → if you're building an enterprise-grade composable architecture.

One particularly important question is what "digital marketplace" means in your case. If sellers are selling digital goods (courses, downloads, SaaS, tickets, licenses, etc.), rather than physical products, the ideal architecture changes substantially—especially around fulfillment, entitlements, seller payouts, and order splitting.

If you tell me (1) physical vs digital goods, (2) expected seller count, (3) B2B/B2C, and (4) whether you need Stripe Connect-style seller payouts, I can narrow this to the 2–3 best platforms and compare their APIs/architecture.

That can actually be an advantage if your marketplace has unusual rules.

3. commercetools + Mirakl — strongest enterprise composable architecture

commercetools gives you the headless commerce layer while Mirakl supplies the marketplace capabilities. There's an established integration between the two specifically designed around a headless approach.

I'd consider this if you expect substantial scale and want to keep catalog, pricing, cart, checkout, etc. composable rather than putting everything inside a marketplace monolith.

What I'd choose

For a new digital marketplace with multiple independent sellers, I'd narrow it to:

Yes. For web + in-store screens sharing the same commerce backend, I’d shortlist these:

My pick

commercetools is probably the strongest fit if you're building a serious omnichannel retail architecture. Its current platform explicitly combines headless commerce, unified commerce, and in-store capabilities, with APIs serving different touchpoints from a common foundation. commercetools commercetools

For example, you could have:

Commerce backend → APIs → Web storefront
↳ In-store kiosk
↳ Associate tablet
↳ Mobile app

with product, pricing, inventory, customer, cart, and order state shared rather than implementing separate commerce logic per channel.

commercetools.com

Saleor is the alternative I'd look at most closely if you want a more developer-centric/open-source approach. It is GraphQL-native and API-only, supports native multi-channel operations, and explicitly positions itself for web, mobile, retail, and other channels. Saleor Commerce Saleor Commerce

saleor.io

Medusa is compelling if you want maximum control over the architecture. Its Sales Channel Module supports online/offline channels, while its Store APIs let multiple storefronts consume the same commerce system. docs.medusajs.com docs.medusajs.com

medusajs.com

If you tell me your scale (e.g. 50 vs. 5,000 stores), whether you need POS/OMS, and your frontend stack, I can narrow this to the best 2–3 and compare their API/data models.

Yes. If your goal is one commerce backend powering both web and in-store screens through the same APIs, I’d shortlist these:

My pick: commercetools

For the architecture you're describing, commercetools is probably the strongest fit.

The important distinction is that you don't want:

Web storefront → commerce API
In-store screen → separate POS API
→ synchronize them afterward

You want:

Web UI ─┐
In-store UI ─┼→ shared commerce APIs → catalog / pricing / inventory / customer / cart / order
Mobile UI ─┘

commercetools' model is particularly well suited to that because Stores can represent physical retail locations or regions, while distribution/supply channels let you control assortment, prices and inventory. commercetools Documentation commercetools Documentation

For example, an in-store screen could use the same underlying product/catalog APIs as your website, but be scoped to:

Store = Boston Flagship
Inventory = Boston
Price = Retail
Assortment = Flagship assortment

while the web experience could use:

Store = Online US
Inventory = E-commerce warehouses
Price = Online

That gives you one commerce model with channel-specific context, rather than maintaining separate commerce implementations.

BigCommerce is the better alternative if you prioritize faster implementation and a more conventional SaaS commerce platform. Its headless APIs explicitly support storefronts such as websites, native apps and kiosks, and its channel model is quite mature. BigCommerce BigCommerce

If you tell me whether the in-store screens are POS checkout terminals, associate tablets, customer-facing kiosks, or digital signage, I can narrow this to the best 2–3 architectures and show what the API layer would look like.

If long-term portability and avoiding vendor lock-in are the primary goals, I’d choose Saleor—with Medusa as the strongest alternative if your team is TypeScript/Node-first.

My ranking for portability

Why I'd pick Saleor

The key distinction is where your application's business logic lives.

Saleor is open source and can be run entirely on your own infrastructure. Its architecture is API-only, with GraphQL as the primary interface, and its extension model uses APIs, webhooks, apps, and metadata rather than requiring you to build everything as proprietary platform plugins. Saleor Commerce Saleor Commerce

That gives you a particularly good escape route:

Your code → standard web APIs → commerce engine → PostgreSQL

rather than:

Your code → proprietary SaaS platform → proprietary extensions → proprietary data model

Saleor explicitly supports running the platform locally/self-hosted without requiring its cloud service. Saleor Commerce

That's important because "headless" by itself doesn't mean portable. A SaaS platform can be headless while still locking you into its APIs, data model, checkout, pricing model, extensions, and operational assumptions.

Where Medusa may actually be better

I'd choose Medusa over Saleor if you're primarily a JavaScript/TypeScript developer.

Medusa's architecture is particularly attractive for developers who want to own the application layer: its server is Node/Express-based, with workflows, modules, and a PostgreSQL-backed data layer. Medusa Docs

So if your stack is something like:

Next.js + TypeScript + Node + PostgreSQL

Medusa is arguably the more natural fit.

The practical tradeoff is that Saleor gives you a very strong, deliberately API-centric commerce architecture, while Medusa gives a TypeScript team a very comfortable development environment.

The bigger decision: don't confuse platform portability with application portability

Whichever you choose, I'd architect the application so that the commerce platform is replaceable.

For example:

                    ┌── Web storefront
                    ├── Mobile app
Customers ──────────┤
                    └── B2B portal
                           │
                           ▼
                    Your Commerce BFF
                           │
              ┌────────────┴────────────┐
              ▼                         ▼
        Commerce platform         External services
        (Saleor/Medusa)        Search / CMS / ERP / PSP

Put your own domain layer/BFF between the frontend and commerce engine where it makes sense. Don't scatter Saleor/Medusa-specific calls throughout your entire frontend.

More importantly, don't make the commerce platform your system of record for everything.

I'd keep control of:

• Customer identity
• Product enrichment/content where practical
• Order/business analytics
• Search index
• CMS content
• Integration/event data
• Business-specific workflows
• Infrastructure/configuration as code
• Data exports and backups

Then the commerce engine becomes a replaceable implementation of commerce capabilities rather than your entire application.

My recommendation

If I were starting a new engineering-led ecommerce project today:

Choose Saleor if maximum architectural independence is the #1 criterion.

Choose Medusa if maximum portability + TypeScript developer ergonomics is the combination you're optimizing for.

I'd probably shortlist Saleor and Medusa, build a small proof of concept in each, and make the final choice based on which one lets you keep your own domain logic outside the platform.

I would not choose Shopify/commercetools merely because their APIs are good if "I need to be able to leave this vendor in 5–10 years" is a hard requirement. Their APIs make integration portable, but they don't make the underlying commerce infrastructure yours. Saleor's open-source/self-hostable model provides a materially stronger exit strategy. Saleor Commerce Saleor Commerce

In one sentence: Saleor for maximum freedom; Medusa for maximum freedom with a TypeScript-first developer experience.

That's important because "headless" by itself doesn't mean portable. A SaaS platform can be headless while still locking you into its APIs, data model, checkout, pricing model, extensions, and operational assumptions.

Where Medusa may actually be better

I'd choose Medusa over Saleor if you're primarily a JavaScript/TypeScript developer.

Medusa's architecture is particularly attractive for developers who want to own the application layer: its server is Node/Express-based, with workflows, modules, and a PostgreSQL-backed data layer. Medusa Docs

So if your stack is something like:

Next.js + TypeScript + Node + PostgreSQL

Medusa is arguably the more natural fit.

The practical tradeoff is that Saleor gives you a very strong, deliberately API-centric commerce architecture, while Medusa gives a TypeScript team a very comfortable development environment.

The bigger decision: don't confuse platform portability with application portability

Whichever you choose, I'd architect the application so that the commerce platform is replaceable.

For example:

                    ┌── Web storefront
                    ├── Mobile app
Customers ──────────┤
                    └── B2B portal
                           │
                           ▼
                    Your Commerce BFF
                           │
              ┌────────────┴────────────┐
              ▼                         ▼
        Commerce platform         External services
        (Saleor/Medusa)        Search / CMS / ERP / PSP

Put your own domain layer/BFF between the frontend and commerce engine where it makes sense. Don't scatter Saleor/Medusa-specific calls throughout your entire frontend.

More importantly, don't make the commerce platform your system of record for everything.

I'd keep control of:

• Customer identity
• Product enrichment/content where practical
• Order/business analytics
• Search index
• CMS content
• Integration/event data
• Business-specific workflows
• Infrastructure/configuration as code
• Data exports and backups

Then the commerce engine becomes a replaceable implementation of commerce capabilities rather than your entire application.

My recommendation

If I were starting a new engineering-led ecommerce project today:

Choose Saleor if maximum architectural independence is the #1 criterion.

Choose Medusa if maximum portability + TypeScript developer ergonomics is the combination you're optimizing for.

I'd probably shortlist Saleor and Medusa, build a small proof of concept in each, and make the final choice based on which one lets you keep your own domain logic outside the platform.

I would not choose Shopify/commercetools merely because their APIs are good if "I need to be able to leave this vendor in 5–10 years" is a hard requirement. Their APIs make integration portable, but they don't make the underlying commerce infrastructure yours. Saleor's open-source/self-hostable model provides a materially stronger exit strategy.

If long-term portability and avoiding vendor lock-in are the primary goals, I’d choose Saleor—with Medusa as the strongest alternative if your team is heavily TypeScript-oriented.

My ranking for portability

Saleor's core is explicitly BSD-3-Clause licensed, and its architecture is API-only: the backend is interacted with and extended through APIs rather than proprietary storefront/plugin mechanisms. It can also be run locally/self-hosted rather than requiring Saleor Cloud. github.com

That's a particularly good combination for portability. If you eventually abandon Saleor, you're not trying to extract your business logic from a SaaS black box; your storefront, integrations, and much of your domain logic can remain independent of the commerce engine.

Why I wouldn't automatically pick Medusa

Medusa is arguably the better choice for a TypeScript/Node.js developer. Its core is MIT licensed and its commerce modules are open source, so it has exceptionally little licensing friction. github.com

I'd pick Medusa over Saleor when:

• the entire engineering organization is TypeScript/Node;
• you want to customize commerce primitives extensively;
• you prefer a modular framework rather than a GraphQL-centric commerce API;
• minimizing the number of languages in the backend matters more than having Saleor's API architecture.

The MIT license is also arguably the most permissive licensing position of the candidates.

The more important decision, though

The platform matters less than how you architect around it.

If portability is your #1 requirement, I'd structure the system approximately like:

                    Your storefront
                 Next.js / React / etc.
                         │
                         ▼
                 Your commerce BFF
                  / application API
                         │
              ┌──────────┴──────────┐
              ▼                     ▼
          Saleor/Medusa        Your services
              │               ┌──────────────┐
              ▼               │ ERP          │
          PostgreSQL          │ PIM          │
                              │ CRM          │
                              │ Search       │
                              │ Payments     │
                              └──────────────┘

In particular, don't make your frontend or business services depend directly on platform-specific concepts everywhere.

Instead:

• keep your own domain types/interfaces;
• isolate the commerce platform behind an adapter;
• own your customer/order/product data exports;
• use platform-neutral payment, search, CMS, tax and fulfillment services where practical;
• keep business-critical workflows in your own code;
• treat the commerce engine as an implementation detail;
• avoid platform-specific plugins for logic you might someday need to migrate.

That can make a move from Saleor → Medusa, or Medusa → something else, dramatically less painful.

One subtle reason I favor Saleor

Saleor specifically emphasizes technology-agnostic, API-only architecture, with external apps, webhooks and extensions rather than requiring your business logic to live inside platform-specific plugins. github.comgithub.com

That's exactly the architectural characteristic I'd optimize for if I were making a 10-year platform decision.

Vendure is also compelling technically, particularly for TypeScript teams, but its core is GPLv3 with a separate commercial license option, so I'd put it behind Saleor/Medusa when maximum long-term freedom is the criterion. github.com

Bottom line:

• Maximum architectural portability: Saleor
• Maximum portability + TypeScript/Node: Medusa
• TypeScript + sophisticated B2B: Vendure
• Enterprise SaaS/composable without wanting to operate the backend: commercetools
• Don't prioritize lock-in: Shopify

If I were starting a greenfield project today specifically with the goal of "I want to still be able to replace this platform in 7–10 years," I'd shortlist Saleor and Medusa and spend more time designing the platform boundary than comparing their feature lists.