JSON to code
Paste a JSON sample and get typed models for Kotlin, C#, Go, Python, Rust or Swift, with renamed fields mapped back correctly.
From a JSON sample to typed models
Every service that consumes a JSON API ends up with the same chore: read an example response and write the classes, structs or data classes for it, in whichever language the service uses. It is mechanical, easy to get subtly wrong (a field that is sometimes missing, a number that is sometimes a decimal) and not interesting. This converter does it for Kotlin, C#, Go, Python, Rust and Swift from a single pasted sample.
Nested objects become their own types, named after the key that holds them, with plural names singularised for list items (orders gives Order). Keys that are not valid identifiers, or that clash with language keywords (type in Rust, class in Kotlin), are renamed and mapped back with the right annotation: @SerialName, JsonPropertyName, a struct tag, a Pydantic alias, #[serde(rename)] or Swift CodingKeys.
Reading the result critically
- A sample shows what was sent, not what can be sent. Merge several samples into one array to teach the tool about optional fields.
- Numbers are modelled as 64-bit integers or doubles. Use decimal types by hand for money.
- Dates arrive as strings. Map them to your language's date type once you know the format.
- Check nullability against the API documentation, not only the sample.
For Java, use the JSON to Java converter on this site, which can also generate records and Lombok classes. Nothing you paste is uploaded.
Frequently asked questions
Which libraries do the generated models target?
Kotlin uses kotlinx.serialization, C# uses System.Text.Json, Go uses encoding/json struct tags, Python offers dataclasses or Pydantic v2, Rust uses serde and Swift uses Codable. Field renames are added where the JSON name differs from the idiomatic name.
How does it decide a field is optional?
A field that is null in the sample, or missing from some objects in an array, becomes nullable or optional (?, Option, omitempty, | None). One sample cannot prove a field is always present, so review the result.
What happens when an array mixes types?
Whole numbers and decimals merge to a floating-point type. Other mixtures, such as strings and numbers in one array, fall back to a generic JSON type (JsonElement, any, Any, serde_json::Value).
Can it also make a JSON Schema?
Yes. Use the JSON Schema generator on this site for draft 2020-12 schemas and OpenAPI components from the same sample.