Standard Schema support

T3 Env is built on the Standard Schema specification, which means you can use any validator library that implements this standard. This gives you the flexibility to choose the validation library that best fits your project's needs, or the one that you're already using.

What is Standard Schema?

Standard Schema is a unified specification for TypeScript schema validation libraries. It provides a common interface that allows different validation libraries to be interoperable. This means that libraries like T3 Env can work with any validator that implements the Standard Schema interface.

The specification defines:

A standardized way to validate input data
Type inference for input and output types
A consistent error format

For more details, visit the Standard Schema specification.

Supported validators

T3 Env supports any validator that implements the Standard Schema v1 interface. Here are some popular options:

Zod

Zod is a TypeScript-first schema validation library with a simple, composable API.

npm install zod

npm install zod

src/env.ts

import { createEnv } from "@t3-oss/env-core";
import { z } from "zod";
 
export const env = createEnv({
  server: {
    DATABASE_URL: z.string().url(),
    PORT: z.coerce.number().default(3000),
  },
  clientPrefix: "PUBLIC_",
  client: {
    PUBLIC_API_URL: z.string().url(),
  },
  runtimeEnv: process.env,
});

src/env.ts

import { createEnv } from "@t3-oss/env-core";
import { z } from "zod";
 
export const env = createEnv({
  server: {
    DATABASE_URL: z.string().url(),
    PORT: z.coerce.number().default(3000),
  },
  clientPrefix: "PUBLIC_",
  client: {
    PUBLIC_API_URL: z.string().url(),
  },
  runtimeEnv: process.env,
});

Valibot

Valibot is a modular, tree-shakeable validation library focused on bundle size and performance.

npm install valibot

npm install valibot

src/env.ts

import { createEnv } from "@t3-oss/env-core";
import * as v from "valibot";
 
export const env = createEnv({
  server: {
    DATABASE_URL: v.pipe(v.string(), v.url()),
    PORT: v.optional(v.pipe(v.string(), v.transform(Number)), "3000"),
  },
  clientPrefix: "PUBLIC_",
  client: {
    PUBLIC_API_URL: v.pipe(v.string(), v.url()),
  },
  runtimeEnv: process.env,
});

src/env.ts

import { createEnv } from "@t3-oss/env-core";
import * as v from "valibot";
 
export const env = createEnv({
  server: {
    DATABASE_URL: v.pipe(v.string(), v.url()),
    PORT: v.optional(v.pipe(v.string(), v.transform(Number)), "3000"),
  },
  clientPrefix: "PUBLIC_",
  client: {
    PUBLIC_API_URL: v.pipe(v.string(), v.url()),
  },
  runtimeEnv: process.env,
});

ArkType

ArkType is a runtime validation library with TypeScript-like syntax and exceptional performance.

npm install arktype

npm install arktype

src/env.ts

import { createEnv } from "@t3-oss/env-core";
import { type } from "arktype";
 
export const env = createEnv({
  server: {
    DATABASE_URL: type("string.url"),
    PORT: type("string.numeric.parse"),
  },
  clientPrefix: "PUBLIC_",
  client: {
    PUBLIC_API_URL: type("string.url"),
  },
  runtimeEnv: process.env,
});

src/env.ts

import { createEnv } from "@t3-oss/env-core";
import { type } from "arktype";
 
export const env = createEnv({
  server: {
    DATABASE_URL: type("string.url"),
    PORT: type("string.numeric.parse"),
  },
  clientPrefix: "PUBLIC_",
  client: {
    PUBLIC_API_URL: type("string.url"),
  },
  runtimeEnv: process.env,
});

Other validators

Any validator that implements the Standard Schema v1 interface will work with T3 Env. Check the Standard Schema ecosystem for more options.

Presets

T3 Env ships with presets for common deployment platforms (Vercel, Netlify, Railway, etc.). These presets are available for multiple validators:

@t3-oss/env-core/presets-zod - Presets using Zod
@t3-oss/env-core/presets-valibot - Presets using Valibot
@t3-oss/env-core/presets-arktype - Presets using ArkType

src/env.ts

import { createEnv } from "@t3-oss/env-core";
import { vercel } from "@t3-oss/env-core/presets-valibot";
import * as v from "valibot";
 
export const env = createEnv({
  server: {
    DATABASE_URL: v.pipe(v.string(), v.url()),
  },
  extends: [vercel()],
  runtimeEnv: process.env,
});

src/env.ts

import { createEnv } from "@t3-oss/env-core";
import { vercel } from "@t3-oss/env-core/presets-valibot";
import * as v from "valibot";
 
export const env = createEnv({
  server: {
    DATABASE_URL: v.pipe(v.string(), v.url()),
  },
  extends: [vercel()],
  runtimeEnv: process.env,
});

Choose the preset import that matches your validator to keep your bundle size optimized.