Proven technology, chosen for the long term
We choose tools that your team can support for years, not the newest trend. When a newer technology is clearly better for the job, we explain why in writing before we use it.
Front end
- React
- Our default choice for interactive interfaces and design systems.
- Next.js
- Fast, search-friendly web applications, with server rendering and image optimisation built in.
- TypeScript
- Used on both the front end and the back end, because catching errors while coding is cheaper than fixing them in production.
- Tailwind CSS
- Consistent, maintainable styling without large, fragile stylesheets.
- HTML and CSS
- Semantic markup underneath everything we build: the foundation of accessibility.
Back end
- Node.js
- APIs and back-end services that share code and types with the front end.
- Python
- Data pipelines, machine learning services and automation.
- .NET
- Enterprise systems and integrations, especially alongside existing Microsoft technology.
- Java
- Long-lived transactional systems in banking and large enterprises.
- PHP and Laravel
- A practical choice for content-driven platforms and for maintaining existing systems.
Mobile
- Flutter
- One codebase for Android and iOS, when budget matters more than a fully native feel.
- React Native
- Cross-platform apps that share code with an existing React web application.
- Swift
- Native iOS apps, when performance or platform features require it.
- Kotlin
- Native Android apps, including support for devices on older versions of Android.
Databases
- MySQL
- Our default relational database: well understood and well supported in Nepal.
- PostgreSQL
- When a project needs richer data types, extensions or more advanced analytical queries.
- MongoDB
- Document storage for data whose structure genuinely varies, rather than as a default.
- Redis
- Caching, rate limiting, queues and session storage.
Cloud platforms
- AWS
- Our most-used cloud platform, including for clients with data residency requirements.
- Microsoft Azure
- Common in organisations that already use Microsoft licensing.
- Google Cloud
- A strong fit for data and machine learning workloads.
DevOps
- Docker
- Identical environments, from a developer’s laptop to production.
- Kubernetes
- Used where the scale justifies it, not on every project.
- GitHub Actions
- Testing and deployment pipelines defined alongside the code.
- Terraform
- Infrastructure as code, so environments can be rebuilt reliably instead of from memory.
AI and machine learning
- OpenAI API
- Language models for assistants, data extraction and content workflows.
- TensorFlow
- Training and serving production models for computer vision and forecasting.
- PyTorch
- Research and experimentation, especially for custom model architectures.
- LangChain
- Orchestration for document-based assistants and multi-step AI workflows.
- scikit-learn
- Classical machine learning models, which are often a better fit than deep learning.
Already have a system we would be taking over?
Share access to the code and infrastructure. We will give you a written assessment of its condition before we quote.