SQL DDL to TypeScript, Zod & Prisma Schema Converter

Modern full-stack applications require end-to-end type safety between database tables, server APIs, and client-side code. Manually writing TypeScript interfaces that mirror database tables is error-prone and leads to subtle runtime bugs when database columns change. Automated schema conversion ensures that SQL column constraints (e.g. NOT NULL, DEFAULT, UUID, VARCHAR limits) are faithfully translated into corresponding TypeScript types, Zod validation guards, and ORM entity models.

The converter maps SQL types to their JavaScript runtime representations: INT/SERIAL/NUMERIC to number / z.number(), UUID/VARCHAR/TEXT to string / z.string(), TIMESTAMPTZ/DATETIME to Date / z.date(), BOOLEAN/BIT to boolean / z.boolean(), and JSONB to Record<string, unknown> / z.record(z.unknown()). Primary keys and NOT NULL columns become required properties, while nullable columns are marked with "| null" and optional "?".

When creating new database records, auto-incrementing integer IDs or default UUIDs (gen_random_uuid()) and timestamp columns (created_at, updated_at) are assigned by the database engine. Our converter generates an accompanying Insert DTO using TypeScript Omit<T, "id" | "created_at">, allowing frontends to validate user forms without requiring auto-generated database fields.

TypeScript types only exist at compile time; they are erased during JavaScript bundling and cannot protect backend endpoints from invalid HTTP payloads. Converting SQL DDL into Zod schemas allows developers to perform runtime parsing (schema.parse(req.body)) while maintaining static types via z.infer<typeof schema>.

Whether using Prisma ORM with declarative prisma.schema models or Drizzle ORM with TypeScript-first pgTable definitions, the converter generates schema definitions with proper @id, @default(), @unique, and @updatedAt attributes.