API design and software architecture
Architecture decides how long software stays easy to change. We design systems around the business domain, give every service a clear contract, and choose the simplest structure that meets the load, from one well-ordered application to a set of services.
The tools
REST and CRUD APIs, OpenAPI contracts, WebSocket, microservices, messaging and domain-driven design.
REST
Resource-based HTTP APIs with consistent naming, status codes, paging and versioning. REST is our default for web, mobile and partner integrations.
CRUD
Create, read, update and delete done properly: validation, permissions, audit trail and safe handling of two people editing the same record. Most business software is CRUD at its core.
OpenAPI
The API is described in an OpenAPI (Swagger) specification first. Frontend, mobile and partner teams get interactive documentation and generated clients, and can build in parallel.
WebSocket
Two-way real-time channels for live dashboards, notifications and multiplayer games, next to the REST API.
Microservices
Independent services, each with its own data and deployment, packaged in containers. We split a system when teams or load call for it, and keep a single well-structured application when they do not.
Event-driven messaging
Services that talk through message queues such as JMS instead of waiting on each other, so a slow or failed step does not stop the rest.
Domain-Driven Design
The code uses the language of the business. Bounded contexts keep each area's rules in one place, which keeps large systems understandable.
Clean architecture
Business rules separated from frameworks, databases and user interface, so each can be tested on its own and replaced without a rewrite.
Reverse proxy
One front door for every application: nginx or Apache terminates TLS, routes each path to the right service, caches static files and blocks unwanted traffic. Every site and product we run sits behind one.
Load balancing
Traffic spread over several application servers with health checks, so load is shared and a server can be taken out for an update without anyone noticing.
Node cluster failover
Clusters in which a healthy node takes over when another fails, with sessions kept alive. We have built clustered application servers for systems that must stay online, such as healthcare.
How we work with them
Contract first
The API contract is agreed and reviewed before code is written, and changes to it are versioned.
Secure by default
OAuth2 and JWT on every endpoint, validated input and rate limits.
Modernizing legacy systems
Old systems are replaced piece by piece behind a stable API, so users keep working during the move.
Documented decisions
Short architecture decision records explain why a choice was made, for the developers who come later.
The experience behind it
Where this technology has been used in real projects.
Explore the rest of the stack
Planning a project on this stack?
Tell us what you are building and we will tell you how we would approach it.
