Quick start

import {
  defineEntity,
  field,
} from "@nexload-sdk/payload-schema";
import type { CollectionConfig } from "payload";

export const productEntity = defineEntity({
  name: "Product",
  fields: {
    title: field.text({ required: true, trim: true }),
    slug: field.slug({ required: true }),
    inventory: field.number({
      integer: true,
      safe: true,
      defaultValue: 0,
    }),
  },
});

export const Products: CollectionConfig = {
  slug: "products",
  fields: productEntity.payload.all(),
};

export const createProductSchema = productEntity.schema(({ pick }) =>
  pick(["title", "slug", "inventory"], {
    optional: ["inventory"],
  }),
);

const input = createProductSchema.parse({
  title: "  New Product  ",
  slug: " New Product ",
});

console.log(input);
// { title: "New Product", slug: "new-product" }

productEntity.payload.all() supplies the fields and canonical write-time normalization to Payload. createProductSchema applies the same field contract to application input before it reaches Payload. inventory is absent above because the picker made it optional; the factory-level defaultValue belongs to Payload and is not converted to Zod .default().

payload.all() preserves declaration order and returns fresh config containers each time. The canonical field adapter is appended after consumer field beforeValidate hooks, so consumer hooks run first and canonical normalization owns the final value.

entity.schema can return any Zod schema:

const titleLength = productEntity.schema(({ fields }) =>
  fields.title.transform((value) => value.length),
);

Do not use the derived schema as a Payload document type. Payload-generated types remain authoritative for persisted documents.

Next, read Concepts for schema ownership and Guides for native fields, relationships, and projections.