Easy to learn
If you've used JSON before, and you've used a type system before, you can learn all of JSON Typedef in just a few minutes.
Type system / IDE integration
Generate TypeScript interfaces, Golang structs, Java classes, and much more from your schemas using the jtd-codegen tool.
Cross-platform consistency
Run input validation anywhere, get conistent results. Every implementation produces the exact same validation errors.
JSON Typedef schemas are written in JSON, in a syntax that's designed to be familiar and unsurprising.
Because schemas are written in JSON, it's easy to parse them yourself for your own custom tooling. You can also easily generate JSON Typedef schemas from other data formats.
// This example uses the JavaScript implementation
// of JSON Typedef.
const jtd = require("jtd");
const schema = {
"properties": {
"name": { "type": "string" },
"created_at": { "type": "timestamp" },
"favorite_numbers": {
"elements": { "type": "float64" }
}
}
}
The example on the right shows how you can use the JavaScript implementation of JSON Typedef. Checking if data is valid is as easy as jtd.validate(schemaJSON, inputJSON).
Check out the docs to see how easy validation is in your preferred programming language.
// Returns an empty array -- no errors!
jtd.validate(schema, {
id: "foo",
created_at: "2020-04-10T20:25:43Z",
favorite_numbers: [1, 3.14, 42]
})
No more useless error: invalid input validation messages. In JSON Typedef, validation errors are part of the spec, and every implementation, across every language, returns the same errors.
Give your users specific errors, without lock-in. With JSON Typedef, you can change what language you do your validation in without changing the validation errors you give your customers.
// Returns three errors:
//
// 1. error at: (root of input)
// (due to: "/properties/id")
//
// 2. error at: "/created_at"
// (due to: "/properties/created_at/type")
//
// 3. error at: "/favorite_numbers/1"
// (due to: "/properties/favorite_numbers/elements/type")
jtd.validate(schema, {
created_at: "yesterday",
favorite_numbers: [1, "3.14", 42]
})
You can generate code in almost any modern programming language from any JSON Typedef schema.
Once you install jtd-codegen, generating code from a schema is as easy as: jtd-codegen --go-out=user/ -- user.jtd.json.
jtd-codegensupports many languages, not just Golang. You can generate code in TypeScript, Java, or Rust, with more languages on the way!// user.jtd.json
{
"properties": {
"name": { "type": "string" },
"created_at": { "type": "timestamp" },
"favorite_numbers": {
"elements": { "type": "float64" }
}
}
}
package user
import "time"
// This is the sort of Go code that jtd-codegen
// would generate from the schema on the left.
type User struct {
Name string `json:"name"`
CreatedAt time.Time `json:"created_at"`
FavoriteNumbers []float64 `json:"favorite_numbers"`
}
Dive into the docs
Learn JSON Typedef
JSON Typedef is designed to be obvious and instantly familiar. See how easy it is to pick up, and then learn how to use JTD in your existing code.
Infer a schema from real data
The jtd-infer tool lets you infer a schema from real data, so you can get introduce JSON Typedef into existing systems in no time.
Create fake data off a schema
With jtd-fuzz, you can generate synthetic data from a JSON Typedef schema. Learn how you can use this to uncover bugs and do load testing.