Digital systems for industry and manufacturing — reliable, integrated, technically honest.

We build the integration with explicit contracts (which entities, in which direction, at what frequency), idempotency, retries, proper error handling and monitoring. Master data, orders, deliveries, prices, stock and status updates are explicit, not implicit. We use standard connectors where they are good enough and build lean middleware where they are not — instead of a plugin pile that breaks on every update.

A first release with login, order history, document downloads, re-orders and RFQ intake typically lands in the mid five-figure range (EUR), depending on the ERP integration and roles model. Expanded portals (with maintenance tickets, service reports, configurators, multi-stage approvals) scale from there. We define the first sensible step after discovery and ship it in 8–14 weeks.

EU hosting, SSO via Azure AD/Entra ID, Keycloak or M365, fine-grained roles, audit trail for all relevant actions, encrypted transport, tested backups, GDPR-compliant logs. Pen tests and code reviews where they make sense. Security is a foundation, not an add-on — and we document it instead of just promising it.

No, if we do it right. Internal tools are built to replace the Excel and paper version without adding new friction. Few clicks, clear shop-floor terminology, good keyboard handling, offline-capable where needed, readable on tablets and PCs. Training is short; a compact guide is enough.

We connect common MES (HYDRA, MPDV, Forcam) as well as custom builds and PLC gateways via OPC UA or MQTT. Typical data points: unit counts, downtime causes, OEE, job status, tooling and material data. Where MES is missing, we suggest a pragmatic first stage (e.g. daily ERP exports plus selective OPC UA polling) instead of demanding a full MES upfront.

Yes. With an explicit redirect plan (URL by URL), preserved critical URLs, a new information architecture along the B2B buyer journey, clean technical SEO, correct hreflang and JSON-LD. Industrial sites often gain visibility after a relaunch because real product content becomes indexable for the first time.

With real localisation, not dictionary translation. German is the leading language; English and any additional languages are maintained with a glossary, technical terminology and market specifics. Dimensions, standards, certificates and country-specific datasheets are modelled in the data layer — not mixed into translation copy.

AI is a tool, not a label. Useful areas: classifying incoming enquiries and drawings, technical translation with a glossary, first-line replies in service tickets with human handover, anomaly alerts on production data, suggestions to complete BOMs. Always with an audit trail and a human in the loop. We do not touch AI on the machine PLCs — that belongs to plant engineers and control specialists.

We build so they can. Standard stack (Next.js, TypeScript, .NET where it fits, PostgreSQL/MS SQL, clear API contracts, documented integrations), traceable architecture decisions, code in the client repository, CI/CD in the client account. Handover is part of the project, not a shock at the end.