JSON Schema Generator

The JSON Schema Generator reads a sample JSON value - an API response, a config file, a webhook payload - and writes a JSON Schema that validates it. It infers the type of every field, merges an array of objects into a single item schema rather than a noisy union, derives which keys are required from the ones present in every record, and can tag string formats like date-time, email, uri and uuid. You paste data you already have; you get the schema you would otherwise hand-write clause by clause.

Writing a schema by hand is slow and easy to get subtly wrong. It is tempting to type what you think a payload contains, miss a nested field, mark an occasionally-null property as required, or forget that one endpoint returns an array where another returns a single object. A schema generated straight from real data starts from the actual shape, so the document you validate against matches what the service genuinely sends. From there you tighten constraints - patterns, enums, minimums - by hand, but the tedious scaffolding is done.

Everything runs in your browser. The inference engine is dependency-free TypeScript, not a wrapper around a hosted API, so payloads - which routinely carry tokens, emails and internal IDs - never leave the page. Choose Draft-07 or the 2020-12 dialect, add a title, toggle required-field detection, format detection and strict `additionalProperties`, then copy the schema or download it as a .json file.

The schema describes your sample rather than your intent, so treat the output as a first draft: required fields are inferred from what happens to be present, and formats are detected by pattern, so a string that looks like a date is typed as one even when it is an identifier that merely resembles one.

Here is a worked example of the JSON schema generator from JSON with more than one record. Paste [{"id":1,"email":"[email protected]","nick":"x"},{"id":2,"email":null}] and press Generate. The output is an array schema whose items have id as integer, email as ["string", "null"] and nick as string, with required set to id and email, because nick is missing from the second record. The header reads 3 properties, 2 required, depth 2, 219 bytes. That is the job of any JSON to JSON schema generator: every record you add can widen a type or drop a key from required.