Technologies We Build With

Our primary mobile stack. Single codebase targeting iOS and Android with native ARM performance. Uses BLoC/Cubit state management and Clean Architecture.

JavaScript-based cross-platform framework for teams with an existing React codebase. Integrates natively with Expo and the React ecosystem.

Native iOS development for projects requiring deep platform integration: ARKit, HealthKit, CoreML, or complex custom animations.

Native Android development for hardware-level integrations, complex background processing, or teams maintaining long-running Android codebases.

Our default web framework. Server components, app router, ISR, and full SSR support. Used for dashboards, marketing sites, and full-scale web applications.

Strict typing across all our web projects. Eliminates a class of runtime errors and improves maintainability as teams and codebases grow.

Scoped component styles with CSS Modules, extended with Sass for variables, mixins, and nesting. No utility-class lock-in.

Build tool for non-Next.js React apps, libraries, and tooling. Sub-second hot reload and optimized production bundles.

Our primary backend stack. Minimal API for lean endpoints, Entity Framework Core for data access, Swagger for documentation. Deployed on Linux containers.

Used for real-time services, lightweight microservices, and webhook receivers where JavaScript runtime consistency with the frontend is preferred.

Used for AI/ML pipelines, data processing scripts, and internal tooling. FastAPI for async HTTP endpoints with automatic OpenAPI schema generation.

We design REST APIs by default and add GraphQL layers when clients need flexible data fetching across multiple resource types.

Default relational database for transactional workloads. Used with EF Core migrations, PostGIS for geospatial data, and pg_trgm for full-text search.

Document store for high-volume event data, audit trails, and use cases where schema flexibility outweighs relational constraints.

In-memory store for caching, rate limiting, session storage, and pub/sub messaging between services.

Embedded database for Flutter offline-first apps (via Drift) and lightweight server-side scenarios where a full database server is overkill.

EC2 for compute, S3 for object storage, Lambda for serverless functions, RDS for managed PostgreSQL. Most production deployments run on AWS.

Used for enterprise clients with existing Microsoft infrastructure. Azure App Service, Azure SQL, and Azure Blob Storage.

Containerized builds for every service. Docker Compose for local dev environments; Docker images promoted through CI/CD to production.

Pipeline automation for builds, tests, Docker image publishing, and deployment triggers. Used across all Codevia-managed projects.

Hosting for frontend web apps with automatic preview deployments per branch. Used for marketing sites and lightweight web applications.

Large language model integration for document analysis, structured data extraction, and AI-assisted workflows. Used via the Messages API and Tool Use.

GPT-4 and embeddings for classification, summarization, and semantic search use cases.

Retrieval-Augmented Generation pipelines that ground LLM outputs in domain-specific document corpora. Paired with pgvector or Pinecone.

Anthropic's open protocol for connecting LLMs to external tools and data sources. Used in internal AI agent tooling and automation workflows.