How to Validate JSON Against a JSON Schema

JSON Schema describes the shape of JSON you expect: required fields, types, enums, nested objects, and arrays. Validating instance documents against a schema catches contract drift before it hits production clients.

JSON Schema basics

A schema is itself a JSON document. Common meta-schema drafts:

• Draft 2020-12 — current $schema for new work
• Draft-07 — still common in OpenAPI 3.0 tooling

Always set "$schema" in your schema file so validators pick the right ruleset.

Core keywords:

• type — string, number, integer, boolean, object, array, null
• properties / required — object fields
• items — array element schema
• enum, const — allowed values
• minimum, maximum, pattern, format — constraints

Example: user profile schema

{
  "$schema": "https://json-schema.org/draft/2020-12/schema",
  "type": "object",
  "required": ["id", "email"],
  "properties": {
    "id": { "type": "string", "format": "uuid" },
    "email": { "type": "string", "format": "email" },
    "role": { "type": "string", "enum": ["member", "admin"] },
    "tags": {
      "type": "array",
      "items": { "type": "string" },
      "maxItems": 10
    }
  },
  "additionalProperties": false
}

additionalProperties: false rejects surprise keys — useful for strict API contracts.

Example: validating an instance

Valid instance:

{
  "id": "550e8400-e29b-41d4-a716-446655440000",
  "email": "dev@example.com",
  "role": "member",
  "tags": ["api"]
}

Invalid instance (missing email, wrong type on role):

{
  "id": 123,
  "role": "superuser"
}

A good validator reports paths to each error: /id expected string, /email required, /role not in enum.

Workflow for API teams

1. Author schema from real response samples or OpenAPI components.schemas.
2. Validate golden fixtures in CI — one valid file per endpoint, plus negative cases.
3. Version schemas alongside API versions (v1/user.schema.json).
4. On failure, diff the instance against last-known-good to see what changed.

OpenAPI and JSON Schema

OpenAPI 3.x embeds JSON Schema (mostly Draft-07 subset) for request/response bodies. The same schema file can power:

• Documentation examples
• Server-side validation (framework-dependent)
• Client code generation

Watch for nullable (OAS 3.0) vs type: ["string", "null"] (plain JSON Schema).

Common validation pitfalls

• integer vs number — 1.0 fails integer type.
• Optional vs required — omitting a key is fine unless listed in required.
• format validation — not all validators enforce format; check your engine.
• Unevaluated properties — Draft 2020-12 uses unevaluatedProperties in combinator cases.

Validate in BracketView

Paste your schema and instance into the JSON Schema Validator. Fix syntax errors in either document with the built-in JSON validator first, then read structured error output with JSON Pointer paths.

See also: JSON Schema glossary term and type generation from JSON.