How do I generate JSON Schema from JSON?

Paste a JSON sample into the tool, select your required field and format options, and click Generate Schema. The tool infers standard JSON Schema Draft 2020-12 syntax locally in your browser.

Can JSON Schema be inferred perfectly from one sample?

A single sample proves that fields exist, but cannot prove whether a missing property might be optional in other payloads. Using multiple samples allows the tool to detect optional fields and nullable types accurately.

How does a generator decide which properties are required?

By default, properties present in every provided sample are marked as required. You can also configure the generator to mark all properties as required or omit the required array entirely.

How are arrays and null values inferred?

Homogeneous arrays produce a single item schema, while heterogeneous array items generate type arrays or anyOf unions. Null values are tracked as the null type.

Can enum, oneOf, anyOf, or allOf be inferred safely from one JSON sample?

Not reliably. One sample cannot prove every allowed enum value or future schema branch, so generated unions and constraints must be reviewed against additional samples and business rules.

Which JSON Schema draft is generated, and how can I validate the result?

The generator outputs JSON Schema Draft 2020-12 with the canonical $schema header. Review the inferred constraints, then validate representative valid and invalid data with a compatible validator such as this site's JSON Schema Validator.

Is my JSON uploaded to a server?

No. All inference, parsing, option evaluation, diagnostics, and sample self-validation run 100% locally in your browser.

Processed locally

JSON to JSON Schema Generator Online

Generate JSON Schema Draft 2020-12 locally from JSON sample(s) with primitive and complex type inference, required field controls, format detection, and diagnostics.

Overview

What this tool does

JSON to JSON Schema Generator infers valid JSON Schema Draft 2020-12 documents locally from JSON sample objects or arrays. It supports single and multi-sample analysis, conservative format detection, configurable required fields and additionalProperties policies, and self-validates samples against the generated schema without network uploads.

Workflow

How to use

Follow the core workflow below. Use “Show in tool” to locate the real control in the live tool above.

Choose one sample or multiple samples
Use Single sample for one representative payload, or open Multiple samples to infer optional fields and shared structure across examples.
Single sampleMultiple samples
Compare multiple samples when fields may be optional
Optional
Open Multiple samples to add representative payloads and distinguish fields that are always present from fields that are optional.
Multiple samples+ Add sample
Add a representative JSON sample
Paste a representative JSON value. In Multiple samples mode, add up to five examples of the same payload shape.
JSON sample+ Add sampleLoad sample
Choose how the schema should be inferred
Set required-field behavior, additional-properties policy, and optional conservative format inference before generating the schema.
Required fieldsAdditional propertiesInfer common string formatsInclude examples
Generate the JSON Schema
Generate a Draft 2020-12 schema locally from the supplied samples.
Generate Schema
Review inference diagnostics and sample validation
Check optional fields, unions, empty structures, inferred formats, and confirm that every provided sample validates against the generated schema.
DiagnosticsSample self-validationWarnings
Copy or download the generated schema
Review the schema, then copy it or download a .schema.json file for validation, tests, or documentation.
Copy SchemaDownload .schema.json

Success

You have generated a standard JSON Schema Draft 2020-12, verified sample self-validation, and exported your schema locally.

Copy SchemaDownload .schema.json

Practical cases

Examples

API response object

Infer property types, required fields, and nested object constraints from a single payload sample.

Optional field across two samples

Compare two samples to automatically detect optional fields when a property is omitted in one payload.

Nullable field

Infer nullable type arrays when a property contains null in one sample and a string in another.

Heterogeneous array items

Generate anyOf or type arrays when array elements contain mixed primitive or object types.

Empty array & object warnings

Receive diagnostic warnings when empty structures leave item or property types underdetermined.

Troubleshooting

Common errors

Assuming an inferred schema automatically discovers hidden business rules like minimum, maximum, or regex patterns.
Treating a single sample as proof that every observed property is permanently required in all API responses.
Over-inferring strict string formats from ambiguous string values.
Generating enum constraints from a single observed string value.
Ignoring empty array and empty object diagnostic warnings when defining target schemas.

From evidence to action

AI workflow

Review the inferred schema first, then use the generated JSON Schema directly in validation, documentation, tests, or a coding-agent workflow. Inference describes the samples you provided; it does not discover hidden business rules.

Developer guides

Learn with this tool

Understand the concept, inspect a testable example, then open it in the tool above.

JSON & SchemaJSON Schema enum: Allowed Values and Examples JSON & SchemaJSON Schema additionalProperties Explained JSON & SchemaJSON Schema oneOf vs anyOf vs allOf

FAQ

JSON to JSON Schema Generator Online

What this tool does

How to use

Choose one sample or multiple samples

Compare multiple samples when fields may be optional

Add a representative JSON sample

Choose how the schema should be inferred

Generate the JSON Schema

Review inference diagnostics and sample validation

Copy or download the generated schema