SwallowKitSchema-driven scaffolding

Keep your Next.js frontend, BFF, Azure Functions backend, and infrastructure aligned through shared Zod schemas.

Shared schema and contracts

Define each domain model once as a Zod schema. SwallowKit uses it to generate TypeScript types, validation, BFF routes, backend handlers, and API contracts — so they stay consistent as the application grows.

Full-stack CRUD scaffolding

Run swallowkit scaffold to generate Azure Functions, Next.js BFF API routes, and React UI components from a single schema file. The generated code is readable and editable.

Azure infrastructure and deployment

Bicep templates, CI/CD workflows (GitHub Actions or Azure Pipelines), and Azure resource provisioning are generated alongside application code. Static Web Apps, Functions, and Cosmos DB are configured with Managed Identity.

AI-agent-friendly structure

Generated instruction files (AGENTS.md, CLAUDE.md, .github/copilot-instructions.md) and the swallowkit machine CLI give coding agents explicit boundaries for inspecting, generating, and validating project artifacts.

Why SwallowKit?

In full-stack applications, the same domain model is repeated across frontend forms, client validation, BFF types, backend DTOs, API contracts, database entities, and infrastructure config. These definitions drift apart as the project grows.

AI tools can generate code quickly, but they do not automatically maintain consistency across these layers. Frameworks can hide complexity, but production applications still need explicit architecture that developers can read and change.

SwallowKit addresses this by placing a shared Zod schema at the center and generating the surrounding application structure from it. It is a scaffolding toolkit — not a runtime framework. Generated code can be read, edited, and replaced.

Next steps

Getting started — Create a project and generate your first CRUD flow
Core concepts — Understand the schema-centric architecture
Scaffold guide — CRUD generation in detail
Deploy to Azure — Provision and deploy with Bicep
AI / MCP guide — Machine-readable interfaces for coding agents