schema

Runtime validation via Standard Schema, the vendor-neutral interface implemented by Zod 3.24+, Valibot, ArkType, and others. schema() turns the <T> cast into a real guarantee.

@itsy/corgi/schema depends only on the Standard Schema type (inlined), so it stays zero-dependency. The actual validator is whatever library you already use.

schema() as a transform

schema(MySchema) returns a transform, use it directly with a request.

import { corgi } from '@itsy/corgi';
import { schema } from '@itsy/corgi/schema';
import { z } from 'zod';

const User = z.object({ id: z.number(), name: z.string() });

// `user` is typed `{ id: number; name: string }` AND validated at runtime:
const user = await corgi.get('/users/1', { transform: schema(User) });

Return type comes from the schema

Because transform drives the return type, no generic is needed — the type comes from the schema.

Errors

A failed validation throws a ValidationError carrying the validator's issues, distinct from HttpError and network errors.

if (isValidationError(err)) {
  for (const issue of err.issues) {
    console.warn(issue.message, issue.path);
  }
}

Any Standard Schema validator works

// Zod, Valibot, ArkType all expose the same `~standard` contract:
import * as v from 'valibot';
import { corgi } from '@itsy/corgi';
import { schema } from '@itsy/corgi/schema';

const User = v.object({ id: v.number(), name: v.string() });
await corgi.get('/users/1', { transform: schema(User) });

parseWith() — the .then() style

If you'd rather not use the transform option.

import { corgi } from '@itsy/corgi';
import { parseWith } from '@itsy/corgi/schema';
import { z } from 'zod';

const User = z.object({ id: z.number(), name: z.string() });

const user = await parseWith(User, corgi.get('/users/1'));