DARINORCold Lab

Tools / JSON Schema Validator

JSON Schema Validator

Validate a JSON document against a JSON Schema right in your browser. Paste a schema and an instance, get every violation with its path — nothing is uploaded.

ResultValid
The instance satisfies the schema.

// About this tool

JSON Schema Validator

This validator checks a JSON document against a JSON Schema entirely in your browser. Paste a schema on one side and an instance on the other, and every violation is reported with the exact path where it failed.

It covers the keywords you actually hit day to day: type, enum, const, required, properties and additionalProperties, items, string bounds (minLength/maxLength/pattern/format), number bounds (minimum/maximum/multipleOf), and composition with allOf/anyOf/oneOf/not.

Everything runs locally — your schema and data are never uploaded, which matters when the payload is a real API request, a config file, or a tool call from an agent.

// When to use it

  • Validate a tool call before it executes

    Check that the arguments an agent is about to pass to a tool actually match the tool's declared schema before the side effect happens.

  • Check an API payload against its contract

    Paste a request or response body and its schema to confirm required fields, types, and formats line up.

  • Debug a config file

    Point a schema at a config's shape and find exactly which field is missing or wrong before it fails at runtime.

// Questions

Is my schema or data uploaded to a server?

No. Validation happens entirely in your browser via JavaScript. Nothing is sent anywhere, which makes it safe for real payloads and internal configs.

Which JSON Schema draft is supported?

A practical subset of draft-07: type, enum, const, required, properties, additionalProperties, items, string/number bounds, pattern, common formats (email, uri, date, date-time, uuid, ipv4), and allOf/anyOf/oneOf/not composition. $ref and remote schema resolution aren't supported.

What does the path in an error mean?

It's the location inside the instance where the violation occurred, written as a slash-separated path — for example /user/email for a nested field.

Can I use this to validate an agent's tool-call arguments?

Yes — that's a common use case. Paste the tool's JSON Schema and the arguments the model produced to confirm they match before the call runs with real side effects.

This site uses minimal cookies and local storage to keep features like the chat widget and games working. We do not use third-party tracking cookies. Privacy Policy