S — Starter
Full-service dedicated
4.900 EUR
- 1 storefront, < 5,000 SKUs
- 1 channel, 1 language
- RAG pipeline, 1 AI use case
- 1 payment, 1 shipping
01 — AI-Ready Commerce as a Service
AI-Ready Commerce as a Service — a shop that AI agents can serve too.
Sylius with an MCP server, RAG pipeline on product data, Schema.org markup as the default, and provenance signatures for manufacturer claims. Composable Commerce on Symfony, ready for SGE, ChatGPT shopping, and customer-support agents that actually know your inventory.
We use Sylius ourselves.
Sylius runs in production at Moselwal itself, not as a demo. What we recommend to you is the platform that carries our own business. AI-Ready Commerce as a Service is the stack we have extracted from that experience and packaged for the mid-market.
Die Realität
Classic shops are built for humans. Not for AI.
With AI-Ready Commerce
- MCP server as the standard interface for AI agents
- Semantic product data automatically rendered as Schema.org JSON-LD
- RAG pipeline on product range, FAQs, and shipping/returns policies
- AI concierge in the storefront optional — customers chat with the shop, which knows stock and order history
- Provenance signature for regulated sectors (food, pharma, cosmetics): manufacturer claims cryptographically verifiable
Until now
- AI search systems only understand products through bog-standard HTML
- No API for AI agents — tool use runs through bespoke endpoints
- Schema.org markup retrofitted as a plug-in, often outdated
- Customer-support bots have no access to real stock, real orders, real delivery times
- Manufacturer claims (organic certification, ingredients, origin) cannot be cryptographically verified
02 — Vier Bausteine
Four building blocks that together make up AI readiness
AI-Ready isn't a switch, it's an architecture. These four components together deliver what classic shop setups simply can't.
Provenance signature
Ed25519 signatures for manufacturer claims, ingredients, and origin information — important for food, cosmetics, pharma, and any sector where product claims must be verifiable.
RAG pipeline on shop data
Indexing of products, FAQs, shipping and returns policies, category descriptions, and manufacturer texts — as the knowledge base for customer-support agents and the product concierge in the storefront.
Semantic product markup
Schema.org JSON-LD for Product, Offer, AggregateRating, Brand, and HowTo generated automatically from the Sylius resource model — no manual template care required.
MCP server for Sylius
Standard tool set for AI agents: product.search, product.get, category.list, order.list, order.get, customer.lookup, inventory.check, cart.get, cart.update, policy.get. Permissions per tool and per agent class.
Two hosting variants, four tiers.
Both variants are fully managed packages including maintenance, updates and major upgrades. Minimum term 12 months; with a monthly contract, onboarding is invoiced once upfront.
Full-Service + Dedicated Hosting
Hosting bei Moselwal in Deutschland, mTLS, CrowdSec WAF, 24/7 Monitoring. Wartung, Updates und Major-Upgrades übernehmen wir, ebenso MCP-Server und RAG-Pipeline. Geeignet, wenn Sie Hosting und Betrieb komplett aus einer Hand wollen.
Full-Service Self-Hosted
Hosting auf Ihrer Infrastruktur, Wartung, Updates und Major-Upgrades übernehmen wir. mTLS, MCP-Server und RAG-Pipeline ebenfalls inkludiert. Geeignet, wenn Sie die Hosting-Hoheit behalten möchten oder regulatorische Vorgaben das fordern.
Full-Service + Dedicated Hosting
Hosting with Moselwal in Germany, mTLS, CrowdSec WAF, 24/7 monitoring. We handle maintenance, updates and major upgrades, including the MCP server and RAG pipeline. Suitable when you want hosting and operations entirely from one provider.
Full-Service Self-Hosted
Hosting on your infrastructure; we handle maintenance, updates and major upgrades. mTLS, MCP server and RAG pipeline included. Suitable when you want to retain hosting sovereignty or when regulations require it.
Hosting variant 1
Full-service + dedicated hosting
M — Growth
Full-service dedicated
7.900 EUR
- 1 storefront, 5–25k SKUs
- 2 channels, 1–2 languages
- RAG + Schema.org Product markup
- 2 AI use cases, 2 payment gateways
- Standard ERP connector
Best Value
L — Scale
Full-service dedicated
11.900 EUR
- 1–2 storefronts, 25–100k SKUs
- Multi-channel, 3–5 languages
- AI customer-support agent as active MCP tool user
- Provenance signatures
- ERP + PIM + tax engine
XL — Enterprise
Full-service dedicated
from
14.900 EUR
- Multi-shop, > 100k SKUs
- B2B order assistant (autonomous pre-filling)
- Custom AI workflows
- GDPR+/EAA/multi-country VAT
- Dedicated solution architect
Hosting variant 2
Full-service self-hosted
S — Starter
Full-service self-hosted
2.900 EUR
- 1 storefront, < 5,000 SKUs
- 1 channel, 1 language
- RAG pipeline, 1 AI use case
- 1 payment, 1 shipping
M — Growth
Full-service self-hosted
4.900 EUR
- 1 storefront, 5–25k SKUs
- 2 channels, 1–2 languages
- RAG + Schema.org Product markup
- 2 AI use cases, 2 payment gateways
- Standard ERP connector
Best Value
L — Scale
Full-service self-hosted
7.900 EUR
- 1–2 storefronts, 25–100k SKUs
- Multi-channel, 3–5 languages
- AI customer-support agent as active MCP tool user
- Provenance signatures
- ERP + PIM + tax engine
XL — Enterprise
Full-service self-hosted
from
9.900 EUR
- Multi-shop, > 100k SKUs
- B2B order assistant (autonomous pre-filling)
- Custom AI workflows
- GDPR+/EAA/multi-country VAT
- Dedicated solution architect
What does BitBag's platform ownership mean for your stack?
Since 2023 Sylius has been owned by BitBag, the platform's largest implementation partner. That is a legitimate market decision, and it is precisely why customer-owned architecture matters more for you, not less. Your code, your data, your stack stay with you, regardless of how the platform's roadmap evolves. Forking readiness included.
What both variants have in common.
Sylius 2 als Fundament
Open Source, Symfony-nativ, API Platform für Headless-Storefronts und Multi-Channel-Anbindungen. Kein Vendor-Lock-in.
MCP-Server für KI-Anbindung
Bestand, Bestellungen und Versand sind als Tools für KI-Agenten verfügbar, mit Audit-Trail je Tool-Aufruf. Standard-konform.
Schema.org + Provenance
Product/Offer/Review-Markup als Standard, kryptografische Signaturen für Hersteller-Aussagen. KI-Suchsysteme finden Ihr Sortiment.
DSGVO-konform
Hosting-Standort Deutschland, AVV in deutsch, Verfahrensverzeichnis als Vorlage, Verantwortliche bleiben Sie.
Sylius 2 as the foundation
Open source, Symfony-native, API Platform for headless storefronts and multi-channel integrations. No vendor lock-in.
MCP server for AI integration
Inventory, orders and shipping are exposed as tools for AI agents, with an audit trail per tool call. Standards-compliant.
Schema.org and provenance
Product/Offer/Review markup as the default, cryptographic signatures for manufacturer claims. AI search systems can find your inventory.
GDPR-compliant
Hosting in Germany, DPA in German, processing register as a template, you remain the data controller.
03 — Use Cases
Three use cases that become realistic right now
These three application scenarios are the obvious first ones — more will emerge as soon as the first productive setups go live.
B2B order assistant
Major customers order by voice command or chat. The agent checks terms (customer-specific pricing, minimum order quantities, delivery SLAs), validates stock, and places orders. Human confirmation optional.
Headless storefront with AI composer
A storefront generator that automatically creates category landing pages, product comparison pages, and theme hubs from product clusters — including Schema.org markup and semantic linking. Editorial decides what goes public.
Customer support agent
Customers chat directly with the shop. The agent knows stock, delivery times, order history, shipping options, and returns rules — real data, not generic boilerplate. Escalation to human staff clearly defined.
04 — Proof aus der Praxis
What we already know from the TYPO3 world
Four posts from our blog that already document the AI-Ready patterns on the TYPO3 side — the architecture transfers to Sylius.
compose.yaml banished from every project
The DevOps setup we'll use to run AI-Ready Commerce in production — the same engineering depth as on the TYPO3 side.
Level 5 Agent-Native: 100/100 in the Cloudflare check
What AI-readiness means on a real platform — demonstrated here on the TYPO3 stack, with the same standard for Sylius.
Content migration as an AI agent
How we run migrations with AI agents — the same mechanics apply to product catalogue migrations.
When AI feeds on your content
What it means when AI agents reach directly into your content — applies in the same way to products and categories.
Composable Commerce on Symfony, AI-ready and without vendor lock-in.
Häufig in Kombination gebucht.
Nächster Schritt
Pilot project or just stay on the list?
If you've got a concrete Sylius use case and want to be a pilot partner: drop us a line. If you just want to stay informed once stage 1 is live: via the engineering newsletter. Both without sales pressure.
Oder einfach informiert bleiben: zum Newsletter · Direkt schreiben: kontakt@moselwal.de
Your path to an AI-Ready shop
Discovery (Woche 1–2)
Bestandsaufnahme, Sortiments-Analyse, Channel-Definition, Tier-Wahl. Am Ende steht eine klare Architektur-Skizze.
Setup (Woche 3–6)
Sylius-Aufbau, MCP-Server-Integration, RAG-Pipeline auf Produktdaten, erster Channel produktiv. Deployment-Pipeline scharf.
AI-Activation (Woche 7–10)
Eval-Datasets, AI-Use-Cases scharfschalten, Schema.org-Markup und Provenance-Signaturen. Erste KI-Tool-Aufrufe gehen produktiv.
Go-Live (Woche 11–12)
Soft-Launch, Monitoring, Übergabe an Ihr Team. Ihre Mitarbeiter:innen führen den Shop danach selbst weiter.
Discovery (week 1–2)
Inventory, assortment analysis, channel definition, tier selection. The result is a clear architecture sketch.
Setup (week 3–6)
Sylius setup, MCP server integration, RAG pipeline on product data, first channel goes live. Deployment pipeline armed.
AI activation (week 7–10)
Eval datasets, AI use cases activated, Schema.org markup and provenance signatures. First AI tool calls go to production.
Go-live (week 11–12)
Soft launch, monitoring, handover to your team. Your staff continues to run the shop without us afterwards.