Integrations and middleware
Two systems, no shared API. That is the part we build.
Custom connectors and queue-based middleware between ERP, CRM, ecommerce, marketplaces, carriers and billing. Designed to survive the day one side goes down.
Integrations and Middleware
Integration work fails in the same three places every time: nobody agreed which field identifies the same record, the webhook arrived twice, and there was no plan for the night a sync failed. We design for those three first and write the API calls afterwards.
Integrations and Middleware
Two systems, no shared API. That is the part we build.
Custom connectors
One-way or two-way, between any pair of systems with an API, a database, an SFTP drop or a file export.
Queue-based middleware
Idempotency keys, retry with exponential backoff, dead-letter queues, and a reconciliation job that reports drift instead of hiding it.
Legacy and EDI adapters
EDIFACT messages, SOAP endpoints, read replicas and portals with nothing but a login. No modern API changes the design, not the feasibility.
Field mapping and data model design
The identity key, the direction of truth, every transformation and the owner of each field, written down before any code exists.
Monitoring and handover
Throughput, failure rate, queue depth and last successful sync, with alerts that name a person and documentation your team can support.
FAQ
How long does one connector take?
Two to six weeks between two systems that both have a documented API. Longer when one side needs a database or SFTP adapter, when historical data has to be migrated, or when the two systems disagree about master data and somebody senior has to decide who wins.
Do you use iPaaS tools or custom code?
Both, chosen per project. n8n and Make suit low-volume flows a client wants to edit without us. Custom middleware wins on volume, complex mapping, or logic that needs tests and version control. We tell you which we recommend before you commit.
What happens when one of the systems goes down?
Messages queue and retry with exponential backoff, so a short outage resolves itself. Anything unrecoverable parks in a dead-letter queue with the payload intact and alerts a named person. Nothing is dropped quietly.
Can you take over an integration somebody else built?
Yes, and we start by documenting what is actually running rather than proposing a rewrite. Around half of these end in a targeted fix, and we say so even though the rebuild would pay us better.
Tell us which two systems are fighting.
One business day, one clear next step. Not a proposal deck.