@drzl/generator-tanstack-start

Generate TanStack Start server functions from a Drizzle schema: one module per table, reads on GET and writes on POST, every payload validated by the schemas DRZL already emits.

The surface that needs no adapting

Measured on 2026-08-11 against @tanstack/react-start 1.168.42. createServerFn().validator(schema) takes any Standard Schema and is properly variance-aware in both directions: the handler receives the schema's output, so a date column's string -> Date transform does real work at the boundary, and the caller supplies its input, so a date crosses the wire as an ISO string and passing a Date from the caller is a compile error. zod, valibot and arktype all behave identically. No adapter, no cast.

What the generator decides

The method. createServerFn defaults to GET, which is right for a read and wrong for every write: a mutation behind a cacheable verb is one an intermediary is entitled to replay.

Install

npm install -D @drzl/generator-tanstack-start

// drzl.config.ts
export default {
  schema: './src/db/schema.ts',
  generators: [
    { kind: 'zod', path: 'src/validators/zod' },
    { kind: 'tanstack-start', path: 'src/server/fns' },
  ],
};

This generator emits no schemas of its own, so the CLI turns on validation.useShared and derives the import path from the sibling generator's path.

One constraint worth knowing about

Start type-checks both ends of a server function for serialisability: the validator's input and the handler's return value. A type of unknown fails either way. That is reachable from a real schema: a customType column with no $type<T>() is one the analyzer cannot type, so it arrives as unknown and Start refuses the whole function. Give the column a type in your Drizzle schema. Date is fine, and so is everything else DRZL emits.

Full documentation: https://drzl.dev/generators/tanstack-start

Licence

Apache-2.0. Generated output is yours under your own project's licence.