Stack
The stack we bet on.
The exact tools Dynamic313 builds with, and why each one earned its place. We change the list when there's a reason to, not when something newer appears.
01 / LANGUAGES
Languages and runtimes
- TypeScript
- Default for everything on the front and most of the back. Shared types between web and mobile are the single biggest reason.
- Node.js
- API and service runtime, so one language spans the whole request path.
- Python
- Data work, model evaluation, and anything touching the ML ecosystem.
- SQL
- Written directly when a query matters. We do not hide the database behind an abstraction that makes it unreadable.
02 / FRONTEND
Frontend and interface
- Next.js
- App Router. Server components by default, client components where interaction genuinely requires them.
- React
- The component model the rest of the stack assumes.
- Tailwind CSS
- Design tokens declared once and consumed as utilities, which is what keeps a design system and its implementation from drifting.
- React Native / Expo
- Mobile, sharing types and API clients with the web codebase.
- Framer Motion
- Used sparingly. Most animation on our own sites is CSS.
03 / DATA
Data and backend
- PostgreSQL
- Primary datastore. Row-level security and pgvector cover more cases than teams expect before reaching for a second system.
- Prisma
- Schema and migrations, with raw SQL where the generated query would be the wrong one.
- Sanity
- Structured content. Portable Text means the same content renders to web, feed, or plain text.
- Stripe
- Billing, plans, and usage metering.
04 / AI
AI and automation
- Claude API
- Primary model API for production features, tool use, and structured extraction.
- OpenAI API
- Used where a specific model or modality fits the task better.
- pgvector
- Embeddings alongside relational data, which avoids operating a separate vector database until scale actually demands one.
- Evaluation harnesses
- Custom, per project. If an AI feature ships to customers, its quality is measured rather than assumed.
05 / INFRASTRUCTURE
Infrastructure and DevOps
- Azure
- Primary cloud, including Static Web Apps, Functions, and managed Postgres.
- Terraform
- Environments as code, so staging and production differ only in the variables passed to them.
- Docker
- Built once and promoted between environments rather than rebuilt per stage.
- GitHub Actions
- CI/CD with a preview environment per pull request.
- Grafana / Sentry
- Metrics and error tracking, wired up before launch rather than after the first incident.
06 / DESIGN
Design and collaboration
- Figma
- Design files that map to tokens in code, not screenshots handed over a wall.
- Storybook
- Component states — including loading, empty, and error — reviewed in isolation.
- Linear
- Issue tracking. Clients get read access to their own project.
- Claude Code
- AI-assisted engineering inside our repositories, with tests as the contract.
Talk to us
Want this stack on your problem?
Send us the rough version. We reply within one business day.
Start a project →