Who should shortlist Shopware in 2026, and who should not
Shortlist Shopware if you run a mid-market store that wants an integrated, API-first commerce feature set without building custom infrastructure, and you do not need native B2B on a free-tier budget. Community Edition and the Rise tier cover product, checkout, content and storefront needs for a single-region or multi-channel store without forcing custom code for basic commerce.
Skip it today if your requirements include rights & roles, budgets & approvals or a quote system and your budget stops below Evolve, which starts at €2,400/month (opens in a new tab). Skip it too if you need to list products for ChatGPT shopping in the EU right now — Shopware’s own product page marks that “coming soon,” not shipped.
Shopware’s core is worth taking seriously on its own terms: Community Edition is MIT-licensed and developed in the open on GitHub (opens in a new tab), a genuine differentiator from source-available or fully closed platforms in this category. That said, w3techs measures Shopware (opens in a new tab) at under 0.1% of all websites and about 0.2% of tracked ecommerce platforms as of July 2026 — this is a specialist choice, not a mass-market default.
This review separates what Shopware has actually shipped from what it announced at Community Day 2026. Treat the two differently when you build a roadmap around either one.
flowchart TD
A["Need native B2B: rights and roles,<br/>budgets and approvals, quotes?"] -->|No| B["Need EU-region ChatGPT<br/>agentic listing today?"]
A -->|Yes| C{"Budget covers Evolve,<br/>from EUR 2400/month?"}
C -->|Yes| D[Shortlist Shopware]
C -->|No| E[Skip: B2B cost exceeds budget]
B -->|No| D
B -->|Yes| F[Skip: EU agentic commerce not available yet]
You should now see: a one-sentence verdict on whether your own B2B, budget and regional requirements clear Shopware’s bar.
Shopware’s edition ladder: what Rise, Evolve and Beyond actually cost and unlock
Above the free Community Edition, Shopware sells three subscription tiers, all prices excluding VAT: Rise from €600/month, Evolve from €2,400/month, and Beyond at custom quote-based pricing (opens in a new tab). Community Edition itself costs nothing and carries no feature ceiling on its own — the ceiling appears once you need something Community Edition and Rise simply cannot do.
The boundary that matters most for a B2B buyer: native B2B Components — rights & roles, budgets & approvals, the quote system — do not exist in Community Edition or Rise at any price. They appear only from Evolve upward. There is no add-on that restores them at a lower tier; the components are absent from the codebase entirely below Evolve.
Total cost of ownership runs well past the licence fee. Budget for implementation-partner time, hosting, and the custom-code cost of building anything Community Edition and Rise cannot do natively — a real line item for any B2B workflow you can’t push onto Evolve.
The common pitfall: a team scopes a B2B rollout against Community Edition, builds for months, then discovers the components don’t exist in that codebase at all. Check the current Evolve and Beyond feature pages before scoping, not general marketing copy or older “B2B Suite” documentation that predates this split.
Every price and boundary in this section was checked in July 2026. Shopware’s pricing page and release notes move the free/paid line by tier and by month, so re-verify before you budget against it.
flowchart LR
subgraph CE["Community Edition - free, MIT"]
C1[Core commerce]
C2[Admin and Store APIs]
end
subgraph RS["Rise - from EUR 600/month"]
R1[Managed hosting]
R2[Support SLA]
end
subgraph EV["Evolve - from EUR 2400/month"]
E1[B2B Components]
E2[Rights and roles]
E3[Budgets and approvals]
E4[Quote system]
end
subgraph BY["Beyond - custom quote"]
B1[Enterprise-scale features]
end
CE --> RS --> EV --> BY
You should now see: the exact tier and monthly price your required B2B workflow forces you onto.
The architecture behind Shopware 6.7: PHP, Symfony, Vue and two release lines
The current stable line is 6.7.x, at 6.7.12.2, released 2026-07-23 (opens in a new tab). Shopware maintains a parallel 6.6.x line at the same time, at 6.6.10.21, released the same day — two branches to track for security patches and plugin compatibility, not one.
The 6.7.0.0 baseline (opens in a new tab), released 2025-06-17, was tested against PHP 8.2 and 8.4, with MySQL 8 or MariaDB 11. Later 6.7.12.x release notes reference PHP 8.2, 8.4 and 8.5, but this pass could not confirm an exact minimum for the patch you’d actually deploy — check that patch’s own release notes before you provision a server.
The Administration is a Vue.js 3 single-page app; the backend runs on a Symfony 7 major version, per Shopware’s own repository description, though the exact 6.7.12.x minor was not independently confirmed. Shopware 6.7 also bundles breaking dependency bumps — PHPUnit to 11.x and Doctrine DBAL to 4.x — that break older test suites and custom data-access code on upgrade from 6.6.x.
One more pitfall to plan around: 6.7.12.0 shipped with a regression that was fixed in 6.7.12.1 and further patched in 6.7.12.2. Pin deployments to 6.7.12.1 or later; do not treat any 6.7.12.x tag as automatically safe.
flowchart TD
subgraph Stack["Shopware 6.7 stack"]
Admin["Administration - Vue.js 3 SPA"] --> API["Store API / Admin API"]
API --> Core["Core - Symfony 7"]
Core --> DB[("MySQL 8 or MariaDB 11")]
end
subgraph Lines["Two maintained lines, same release date"]
L67["6.7.x - current stable - 6.7.12.2"]
L66["6.6.x - parallel maintenance - 6.6.10.21"]
end
You should now see: which PHP, database and dependency versions to pin, and which specific patch level to avoid.
Where Shopware’s open-core model earns its keep
The Community Edition core is genuinely MIT-licensed (opens in a new tab), not source-available and not copyleft. You can fork it, self-host it or resell it without signing a commercial agreement — a real distinction from platforms that call themselves “open” while gating the code behind a licence check.
Development happens in the open on GitHub, so you can inspect commit activity, open issues and release cadence directly instead of trusting a roadmap slide. That’s how this review confirmed the 6.7.12.0 regression and its fix timeline, rather than taking Shopware’s word for it.
Shopware’s own store describes more than 3,100 community extensions (opens in a new tab) across B2B, ERP/PIM/CRM connectors, payments, storefront customisation and AI tools. Treat that figure as medium confidence — no independently confirmed running total exists outside Shopware’s own count.
The Administration and Store APIs are documented and API-first, giving integrators a real surface to build against instead of reverse-engineering the admin UI. That matters more as agentic tooling starts calling those same APIs directly.
You should now see: the specific, verifiable reasons Shopware’s core beats a fully closed competitor on inspectability and extension supply, independent of price.
Where the fit weakens: edition boundaries and regional gaps
The B2B Components ceiling — Evolve upward, from €2,400/month — is the single largest edition-boundary cost this review found. If your commerce model requires quotes, budgets or role-based purchasing, that cost is not optional; it’s the price of entry.
Shopware’s Agentic Commerce sales channel (opens in a new tab) — an AI-agent-readable product feed for platforms including ChatGPT, built on a Universal Commerce Protocol — is US-only today. Shopware’s own product page lists European availability as “coming soon,” with no committed date.
The announcements from Community Day 2026 (opens in a new tab) — Shopware Nexus, an agentic Copilot, native Shopware Payments, Experience Studio — sit at various stages of preview or regional limitation. None is a generally available, purchasable feature today, whatever the stage demo implied.
Shopware Payments itself illustrates the pattern: it launched only in Germany and Austria, with EU and US rollout merely planned. Shopware’s newest products tend to ship regionally narrow first and expand later — plan around the shipped region, not the press release.
You should now see: a clear separation between what Shopware has actually shipped and where, versus what it announced on stage.
Scoring Shopware’s modernity and AI/agent readiness against the CodeNx rubric
Modernity score: 7/10. Shopware earns credit for a maintained runtime and dependency posture — PHP 8.2+/8.4 support and an actively patched 6.7.x line — and for API-first boundaries backed by documented Admin and Store APIs. Its extension model, distributed through the Shopware Store rather than core modification, and its visible automated-test and CI activity on GitHub both count in its favour.
Points come off the modernity score for release-activity and migration-clarity: the 6.7.12.0 regression, fixed in 6.7.12.1 and patched again in 6.7.12.2, happened inside a single point-release cycle. A platform that ships a regression and two follow-up patches in the same month is maintained, but not yet disciplined about it.
AI/agent readiness score: 6/10. An ecommerce platform isn’t AI-ready merely because a chatbot plugin exists somewhere in its store — Shopware clears a higher bar than that with an official, first-party MCP server (opens in a new tab), shipped experimentally from release 6.7.11.0 (2026-06-16) behind the MCP_SERVER feature flag. That’s real credit under “official AI, MCP, agent tooling,” not a bolted-on integration.
The MCP server’s design maps directly onto two more rubric components: its ACL enforcement and configurable allowlists satisfy “safe authenticated write operations,” and its dry-run validation for write operations satisfies “auditability, idempotency, approval-friendly operations.” Both are details a rubric rewards and a press release wouldn’t mention.
Points come off for reach and default state. The Agentic Commerce channel’s (opens in a new tab) US-only availability limits the “product/search data suitable for retrieval” component’s reach for merchants concentrated in the DACH region — a concentration commonly cited for Shopware’s install base but not independently measured here — and the feature-flag gating means the MCP server isn’t on by default for any existing store.
Treat Community Day 2026’s (opens in a new tab) own claims — a stated 15x increase in AI-driven merchant traffic, and “100M+ ChatGPT users already shop” — as marketing figures with no disclosed methodology. Neither feeds either score above; both would need an independent measurement this review doesn’t have.
Production concern: the upgrade risk a Shopware feature table never shows you
A feature-comparison table tells you what a release does. It says nothing about what happens when a point release regresses, and 6.7.12.0 shipped with a breaking issue (opens in a new tab) that was fixed in 6.7.12.1 and further patched in 6.7.12.2 — all inside the same July 2026 release cycle.
Running two simultaneously maintained lines compounds the risk. With 6.7.x and 6.6.x both active, a team has to decide, and keep deciding, which line its custom plugins and CI target. There isn’t a single “current version” to assume; there are two, moving on independent patch schedules.
The PHPUnit 11 and Doctrine DBAL 4 bumps bundled into 6.7 are breaking changes for existing test suites and custom data-access code. An upgrade that looks like a routine point-release bump on a changelog can fail CI or silently change query behaviour once it hits a real plugin.
The lesson generalises past Shopware: on any platform with major dependency bumps between minor versions, run the official upgrade guide against a staging environment with representative plugin coverage before touching production data. Pin to the first patch release known to fix a regression, never to the release that introduced it.
sequenceDiagram
participant Shopware
participant Store as Production store
Shopware->>Store: 6.7.12.0 released - regression introduced
Store->>Store: Upgrade applied, regression hits
Shopware->>Store: 6.7.12.1 released - regression fixed
Shopware->>Store: 6.7.12.2 released - further patch, same cycle
Note over Store: Pin to 6.7.12.1 or later, never 6.7.12.0
You should now see: the reason to budget a staging-environment rehearsal and a specific patch-level pin before any Shopware upgrade, not just a version-number bump.
The five-question proof-of-concept checklist before you commit budget to Shopware
Run these five checks against your own trial account and codebase before signing anything.
1. Does your required B2B workflow actually exist in the tier you’re budgeting for? Check rights & roles, budgets & approvals and the quote system against the current Evolve and Beyond feature pages, not general marketing copy or older documentation.
2. What PHP version does your target 6.7.12.x patch actually require? Confirm it against that patch’s own release notes on GitHub, not the 6.7.0.0 baseline notes, which cover an earlier point in the line.
3. Is the MCP_SERVER feature flag enabled on your instance?
curl -s -o /dev/null -w "%{http_code}\n" https://your-shop.example/api/_mcpA 404 means the flag isn’t enabled or the route isn’t registered on your version. The MCP server enforces its own ACL layer (opens in a new tab), so an unauthenticated request may also return a 401 or 403 rather than a tool list — check with a valid API credential before concluding the feature is absent.
4. Is your merchant region eligible for the Agentic Commerce sales channel today? Check Shopware’s Agentic Commerce product page (opens in a new tab) directly rather than assuming eligibility from the press release — it is the source that actually states US-only, EU “coming soon.”
5. Does your team have a staging environment that can run the official 6.6-to-6.7 upgrade guide against your actual plugin set? If the answer is no, budget the time to build one before this touches production, not after.
Limitations
This review did not include an independent hands-on test of the MCP server’s tool surface, its allowlist behaviour or its dry-run validation in practice — the description above comes from Shopware’s own documentation, not from calling the endpoint. Any DACH-concentration figures referenced for Shopware’s install base are third-party estimates, not measured facts, and are not asserted as such here. Every price and edition boundary in this piece is a dated snapshot from July 2026; Shopware’s pricing page and release notes have moved this boundary before and will again, so re-check both directly before you sign anything.
You should now see: five yes/no checks to run against your own Shopware trial or sales call, and an explicit account of what this review did not verify.