valibot

SKILL.md

Valibot

This skill helps you work effectively with Valibot, the modular and type-safe schema library for validating structural data.

When to use this skill

When the user asks about schema validation with Valibot

When creating or modifying Valibot schemas

When parsing or validating user input

When the user mentions Valibot, schema, or validation

When migrating from Zod to Valibot

CRITICAL: Valibot vs Zod — Do Not Confuse!

Valibot and Zod have different APIs. Never mix them up!

Key Differences

Feature
Zod ❌
Valibot ✅

Import
import { z } from 'zod'
import * as v from 'valibot'

Validations
Chained methods: .email().min(5)
Pipeline: v.pipe(v.string(), v.email(), v.minLength(5))

Parsing
schema.parse(data)
v.parse(schema, data)

Safe parsing
schema.safeParse(data)
v.safeParse(schema, data)

Optional
z.string().optional()
v.optional(v.string())

Nullable
z.string().nullable()
v.nullable(v.string())

Default
z.string().default('x')
v.optional(v.string(), 'x')

Transform
z.string().transform(fn)
v.pipe(v.string(), v.transform(fn))

Refine/Check
z.string().refine(fn)
v.pipe(v.string(), v.check(fn))

Enum
z.enum(['a', 'b'])
v.picklist(['a', 'b'])

Native enum
z.nativeEnum(MyEnum)
v.enum(MyEnum)

Union
z.union([a, b])
v.union([a, b])

Discriminated union
z.discriminatedUnion('type', [...])
v.variant('type', [...])

Intersection
z.intersection(a, b)
v.intersect([a, b])

Min/max length
.min(5).max(10)
v.minLength(5), v.maxLength(10)

Min/max value
.gte(5).lte(10)
v.minValue(5), v.maxValue(10)

Infer type
z.infer<typeof Schema>
v.InferOutput<typeof Schema>

Infer input
z.input<typeof Schema>
v.InferInput<typeof Schema>

Common Mistakes to Avoid

// ❌ WRONG - This is Zod syntax, NOT Valibot!
const Schema = v.string().email().min(5);
const result = Schema.parse(data);

// ✅ CORRECT - Valibot uses functions and pipelines
const Schema = v.pipe(v.string(), v.email(), v.minLength(5));
const result = v.parse(Schema, data);

// ❌ WRONG - Zod-style optional
const Schema = v.object({
  name: v.string().optional(),
});

// ✅ CORRECT - Valibot wraps with optional()
const Schema = v.object({
  name: v.optional(v.string()),
});

// ❌ WRONG - Zod-style default
const Schema = v.string().default("hello");

// ✅ CORRECT - Valibot uses second argument
const Schema = v.optional(v.string(), "hello");

Installation

npm install valibot     # npm
yarn add valibot        # yarn
pnpm add valibot        # pnpm
bun add valibot         # bun

Import with a wildcard (recommended):

import * as v from "valibot";

Or with individual imports:

import { object, string, pipe, email, parse } from "valibot";

Mental Model

Valibot's API is divided into three main concepts:

1. Schemas

Schemas define the expected data type. They are the starting point.

import * as v from "valibot";

// Primitive schemas
const StringSchema = v.string();
const NumberSchema = v.number();
const BooleanSchema = v.boolean();
const DateSchema = v.date();

// Complex schemas
const ArraySchema = v.array(v.string());
const ObjectSchema = v.object({
  name: v.string(),
  age: v.number(),
});

2. Methods

Methods help you use or modify schemas. The schema is always the first argument.

// Parsing
const result = v.parse(StringSchema, "hello");
const safeResult = v.safeParse(StringSchema, "hello");

// Type guard
if (v.is(StringSchema, data)) {
  // data is typed as string
}

3. Actions

Actions validate or transform data within a pipe(). They MUST be used inside pipelines.

// Actions are used in pipe()
const EmailSchema = v.pipe(
  v.string(),
  v.trim(),
  v.email(),
  v.endsWith("@example.com"),
);

Pipelines

Pipelines extend schemas with validation and transformation actions. A pipeline always starts with a schema, followed by actions.

import * as v from "valibot";

const UsernameSchema = v.pipe(
  v.string(),
  v.trim(),
  v.minLength(3, "Username must be at least 3 characters"),
  v.maxLength(20, "Username must be at most 20 characters"),
  v.regex(
    /^[a-z0-9_]+$/i,
    "Username can only contain letters, numbers, and underscores",
  ),
);

const AgeSchema = v.pipe(
  v.number(),
  v.integer("Age must be a whole number"),
  v.minValue(0, "Age cannot be negative"),
  v.maxValue(150, "Age cannot exceed 150"),
);

Common Validation Actions

String validations:

v.email() — Valid email format

v.url() — Valid URL format

v.uuid() — Valid UUID format

v.regex(pattern) — Match regex pattern

v.minLength(n) — Minimum length

v.maxLength(n) — Maximum length

v.length(n) — Exact length

v.nonEmpty() — Not empty string

v.startsWith(str) — Starts with string

v.endsWith(str) — Ends with string

v.includes(str) — Contains string

Number validations:

v.minValue(n) — Minimum value (>=)

v.maxValue(n) — Maximum value (<=)

v.gtValue(n) — Greater than (>)

v.ltValue(n) — Less than (<)

v.integer() — Must be integer

v.finite() — Must be finite

v.safeInteger() — Safe integer range

v.multipleOf(n) — Must be multiple of n

Array validations:

v.minLength(n) — Minimum items

v.maxLength(n) — Maximum items

v.length(n) — Exact item count

v.nonEmpty() — At least one item

v.includes(item) — Contains item

v.excludes(item) — Does not contain item

Custom Validation with check()

const PasswordSchema = v.pipe(
  v.string(),
  v.minLength(8),
  v.check(
    (input) => /[A-Z]/.test(input),
    "Password must contain an uppercase letter",
  ),
  v.check((input) => /[0-9]/.test(input), "Password must contain a number"),
);

Value Transformations

These actions modify the value without changing its type:

String transformations:

v.trim() — Remove leading/trailing whitespace

v.trimStart() — Remove leading whitespace

v.trimEnd() — Remove trailing whitespace

v.toLowerCase() — Convert to lowercase

v.toUpperCase() — Convert to uppercase

Number transformations:

v.toMinValue(n) — Clamp to minimum value (if less than n, set to n)

v.toMaxValue(n) — Clamp to maximum value (if greater than n, set to n)

const NormalizedEmailSchema = v.pipe(
  v.string(),
  v.trim(),
  v.toLowerCase(),
  v.email(),
);

// Clamp number to range 0-100
const PercentageSchema = v.pipe(v.number(), v.toMinValue(0), v.toMaxValue(100));

Type Transformations

For converting between data types, use these built-in transformation actions:

v.toNumber() — Convert to number

v.toString() — Convert to string

v.toBoolean() — Convert to boolean

v.toBigint() — Convert to bigint

v.toDate() — Convert to Date

// Convert string to number
const PortSchema = v.pipe(v.string(), v.toNumber(), v.integer(), v.minValue(1));

// Convert ISO string to Date
const TimestampSchema = v.pipe(v.string(), v.isoDateTime(), v.toDate());

// Convert to boolean
const FlagSchema = v.pipe(v.string(), v.toBoolean());

Custom Transformations

For custom transformations, use v.transform():

const DateStringSchema = v.pipe(
  v.string(),
  v.isoDate(),
  v.transform((input) => new Date(input)),
);

// Custom object transformation
const UserSchema = v.pipe(
  v.object({
    firstName: v.string(),
    lastName: v.string(),
  }),
  v.transform((input) => ({
    ...input,
    fullName: `${input.firstName} ${input.lastName}`,
  })),
);

Object Schemas

Basic Object

const UserSchema = v.object({
  id: v.number(),
  name: v.string(),
  email: v.pipe(v.string(), v.email()),
  age: v.optional(v.number()),
});

type User = v.InferOutput<typeof UserSchema>;

Object Variants

// Regular object - strips unknown keys (default)
const ObjectSchema = v.object({ key: v.string() });

// Loose object - allows and preserves unknown keys
const LooseObjectSchema = v.looseObject({ key: v.string() });

// Strict object - throws on unknown keys
const StrictObjectSchema = v.strictObject({ key: v.string() });

// Object with rest - validates unknown keys against a schema
const ObjectWithRestSchema = v.objectWithRest(
  { key: v.string() },
  v.number(), // unknown keys must be numbers
);

Optional and Nullable Fields

const ProfileSchema = v.object({
  // Required
  name: v.string(),

  // Optional (can be undefined or missing)
  nickname: v.optional(v.string()),

  // Optional with default
  role: v.optional(v.string(), "user"),

  // Nullable (can be null)
  avatar: v.nullable(v.string()),

  // Nullish (can be null or undefined)
  bio: v.nullish(v.string()),

  // Nullish with default
  theme: v.nullish(v.string(), "light"),
});

Object Methods

const BaseSchema = v.object({
  id: v.number(),
  name: v.string(),
  email: v.string(),
  password: v.string(),
});

// Pick specific keys
const PublicUserSchema = v.pick(BaseSchema, ["id", "name"]);

// Omit specific keys
const UserWithoutPasswordSchema = v.omit(BaseSchema, ["password"]);

// Make all optional
const PartialUserSchema = v.partial(BaseSchema);

// Make all required
const RequiredUserSchema = v.required(PartialUserSchema);

// Merge objects
const ExtendedUserSchema = v.object({
  ...BaseSchema.entries,
  createdAt: v.date(),
});

Cross-Field Validation

const RegistrationSchema = v.pipe(
  v.object({
    password: v.pipe(v.string(), v.minLength(8)),
    confirmPassword: v.string(),
  }),
  v.forward(
    v.partialCheck(
      [["password"], ["confirmPassword"]],
      (input) => input.password === input.confirmPassword,
      "Passwords do not match",
    ),
    ["confirmPassword"],
  ),
);

Arrays and Tuples

Arrays

const TagsSchema = v.pipe(
  v.array(v.string()),
  v.minLength(1, "At least one tag required"),
  v.maxLength(10, "Maximum 10 tags allowed"),
);

// Array of objects
const UsersSchema = v.array(
  v.object({
    id: v.number(),
    name: v.string(),
  }),
);

Tuples

// Fixed-length array with specific types
const CoordinatesSchema = v.tuple([v.number(), v.number()]);
// Type: [number, number]

// Tuple with rest
const ArgsSchema = v.tupleWithRest(
  [v.string()], // first arg is string
  v.number(), // rest are numbers
);
// Type: [string, ...number[]]

Unions and Variants

Union

const StringOrNumberSchema = v.union([v.string(), v.number()]);

const StatusSchema = v.union([
  v.literal("pending"),
  v.literal("active"),
  v.literal("inactive"),
]);

Picklist (for string/number literals)

// Simpler than union of literals
const StatusSchema = v.picklist(["pending", "active", "inactive"]);

const PrioritySchema = v.picklist([1, 2, 3]);

Variant (discriminated union)

Use variant for better performance with discriminated unions:

const EventSchema = v.variant("type", [
  v.object({
    type: v.literal("click"),
    x: v.number(),
    y: v.number(),
  }),
  v.object({
    type: v.literal("keypress"),
    key: v.string(),
  }),
  v.object({
    type: v.literal("scroll"),
    direction: v.picklist(["up", "down"]),
  }),
]);

Parsing Data

parse() — Throws on Error

import * as v from "valibot";

const EmailSchema = v.pipe(v.string(), v.email());

try {
  const email = v.parse(EmailSchema, "jane@example.com");
  console.log(email); // 'jane@example.com'
} catch (error) {
  console.error(error); // ValiError
}

safeParse() — Returns Result Object

const result = v.safeParse(EmailSchema, input);

if (result.success) {
  console.log(result.output); // Valid data
} else {
  console.log(result.issues); // Array of issues
}

is() — Type Guard

if (v.is(EmailSchema, input)) {
  // input is typed as string
}

Configuration Options

// Abort early - stop at first error
v.parse(Schema, data, { abortEarly: true });

// Abort pipe early - stop pipeline at first error
v.parse(Schema, data, { abortPipeEarly: true });

Type Inference

import * as v from "valibot";

const UserSchema = v.object({
  name: v.string(),
  age: v.pipe(v.string(), v.transform(Number)),
  role: v.optional(v.string(), "user"),
});

// Output type (after transformations and defaults)
type User = v.InferOutput<typeof UserSchema>;
// { name: string; age: number; role: string }

// Input type (before transformations)
type UserInput = v.InferInput<typeof UserSchema>;
// { name: string; age: string; role?: string | undefined }

// Issue type
type UserIssue = v.InferIssue<typeof UserSchema>;

Error Handling

Custom Error Messages

const LoginSchema = v.object({
  email: v.pipe(
    v.string("Email must be a string"),
    v.nonEmpty("Please enter your email"),
    v.email("Invalid email format"),
  ),
  password: v.pipe(
    v.string("Password must be a string"),
    v.nonEmpty("Please enter your password"),
    v.minLength(8, "Password must be at least 8 characters"),
  ),
});

Flattening Errors

const result = v.safeParse(LoginSchema, data);

if (!result.success) {
  const flat = v.flatten(result.issues);
  // { nested: { email: ['Invalid email format'], password: ['...'] } }
}

Issue Structure

Each issue contains:

kind: 'schema' | 'validation' | 'transformation'

type: Function name (e.g., 'string', 'email', 'min_length')

input: The problematic input

expected: What was expected

received: What was received

message: Human-readable message

path: Array of path items for nested issues

Fallback Values

// Static fallback
const NumberSchema = v.fallback(v.number(), 0);
v.parse(NumberSchema, "invalid"); // Returns 0

// Dynamic fallback
const DateSchema = v.fallback(v.date(), () => new Date());

Recursive Schemas

import * as v from "valibot";

type TreeNode = {
  value: string;
  children: TreeNode[];
};

const TreeNodeSchema: v.GenericSchema<TreeNode> = v.object({
  value: v.string(),
  children: v.lazy(() => v.array(TreeNodeSchema)),
});

Async Validation

For async operations (e.g., database checks), use async variants:

import * as v from "valibot";

const isUsernameAvailable = async (username: string) => {
  // Check database
  return true;
};

const UsernameSchema = v.pipeAsync(
  v.string(),
  v.minLength(3),
  v.checkAsync(isUsernameAvailable, "Username is already taken"),
);

// Must use parseAsync
const username = await v.parseAsync(UsernameSchema, "john");

JSON Schema Conversion

import { toJsonSchema } from "@valibot/to-json-schema";
import * as v from "valibot";

const EmailSchema = v.pipe(v.string(), v.email());
const jsonSchema = toJsonSchema(EmailSchema);
// { type: 'string', format: 'email' }

Naming Conventions

Convention 1: Same Name (Recommended for simplicity)

export const User = v.object({
  name: v.string(),
  email: v.pipe(v.string(), v.email()),
});

export type User = v.InferOutput<typeof User>;

// Usage
const users: User[] = [];
users.push(v.parse(User, data));

Convention 2: With Suffixes (Recommended when input/output differ)

export const UserSchema = v.object({
  name: v.string(),
  age: v.pipe(v.string(), v.transform(Number)),
});

export type UserInput = v.InferInput<typeof UserSchema>;
export type UserOutput = v.InferOutput<typeof UserSchema>;

Common Patterns

Login Form

const LoginSchema = v.object({
  email: v.pipe(
    v.string(),
    v.nonEmpty("Please enter your email"),
    v.email("Invalid email address"),
  ),
  password: v.pipe(
    v.string(),
    v.nonEmpty("Please enter your password"),
    v.minLength(8, "Password must be at least 8 characters"),
  ),
});

API Response

const ApiResponseSchema = v.variant("status", [
  v.object({
    status: v.literal("success"),
    data: v.unknown(),
  }),
  v.object({
    status: v.literal("error"),
    error: v.object({
      code: v.string(),
      message: v.string(),
    }),
  }),
]);

Environment Variables

const EnvSchema = v.object({
  NODE_ENV: v.picklist(["development", "production", "test"]),
  PORT: v.pipe(v.string(), v.transform(Number), v.integer(), v.minValue(1)),
  DATABASE_URL: v.pipe(v.string(), v.url()),
  API_KEY: v.pipe(v.string(), v.minLength(32)),
});

const env = v.parse(EnvSchema, process.env);

Date Handling

// String to Date
const DateFromStringSchema = v.pipe(
  v.string(),
  v.isoDate(),
  v.transform((input) => new Date(input)),
);

// Date validation
const FutureDateSchema = v.pipe(
  v.date(),
  v.minValue(new Date(), "Date must be in the future"),
);

Additional Resources

Valibot Documentation

Valibot GitHub

API Reference

Migration from Zod