01
Include different records
Place several objects of the same response shape in a top-level array. A key appearing only later becomes nullable, while nested arrays are inspected across their elements.
Type generation
JSON to modelfrom one sample to five source files
Try the two-record example, then choose a target language and a root type name.
JSON to model
JSON input
The JSON is parsed and generated in this browser; no sample is uploaded or stored.
Generated model · Rust
2 declarations · 2 nullable types
JSON to model turns a representative API response into a starting type. One array can show both records with different keys, so a field found only in the second record is not lost.
Use the generated declaration as a draft. A single response cannot establish business rules, future fields, exact decimal behavior or your project's serializer configuration.
Why use JSON to model?
Inspect every array record
JSON to model reads all objects in the top-level array. Fields missing from some records remain in the generated type as nullable rather than disappearing because the first record lacked them.
Switch the destination, keep the sample
The same local JSON becomes Rust structs, C# classes, Dart classes, PHP classes or Scala case classes. Each download carries the selected language's file extension.
Make the sample prove the shape
02
Choose the output contract
Rust emits Serde fields; C# uses System.Text.Json attributes; Dart includes fromJson and toJson; PHP and Scala emit model declarations that still need project-side hydration or codecs.
03
Export only the current result
Auto-generation waits 350 ms after edits. Turn it off to reveal Generate model or press Ctrl/Cmd+Enter. Any input or option change disables copy and download until the new run succeeds.
Five concrete model outputs
Rust and C# mappings
A JSON to Rust struct uses Serde rename attributes, Option and Vec. C# retains original keys with JsonPropertyName and uses nullable properties where the sample shows gaps.
Dart model constructors
The Dart model generator emits typed fields plus fromJson and toJson methods for nested objects and lists. Test generated code with your actual Flutter or Dart runtime.
PHP and Scala declarations
PHP 8 receives typed properties and array item docblocks. A Scala case class receives Option and List where evidence supports them; a codec is still a separate project choice.
Lossless inference step
This JSON to class generator reads number tokens before JavaScript can round them. Out-of-range or decimal values fall back to broad target-language types instead of a falsely precise integer.
What sample inference cannot know
- Only five targets are offered
- The selector contains Rust, C#, Dart, PHP and Scala. TypeScript, Go, Java, Python and Swift have their own tools; unavailable languages do not appear as working choices here.
- A sample is not a schema
- JSON Schema, OpenAPI and URL fetching are outside this JSON to model workflow. Required flags, ranges, enum members, dates and currency semantics cannot be proved from one sample.
- Unknown numbers need review
- Integer tokens inside signed 64-bit range get a narrow integer type. Larger integers and decimals get broad types; the generated application may still need a serializer setting for exact numeric round trips.
- A declaration is not a complete application
- PHP properties do not hydrate themselves from json_decode, and Scala case classes need a codec. Rust requires Serde dependencies. Generated code is not executed or compiled on this page.
- Strict JSON and bounded output
- Comments, duplicate keys, trailing commas and mixed top-level array values are rejected. The parser limits input to 500,000 characters, depth 100 and 50,000 nodes; output is capped at 1,000,000 characters.
JSON to model questions
Which target languages can the generator produce?
JSON to model supports Rust, C#, Dart, PHP and Scala on this page. The selector lists only these implemented outputs; each has a matching source-file extension.
How are null and missing fields represented?
All objects in the top-level array are examined. If a key is missing in some records or its value is null, the field becomes nullable where the target language supports that. A null-only field remains an unknown type.
Are nested objects and arrays supported?
Yes. Nested objects become named models, and arrays infer an item type from all visible elements. Mixed item kinds fall back to broad types that you should refine against more examples.
Can I rename the root model?
Yes. Root is the default. Enter an ASCII name starting with a capital letter; changing it regenerates the declarations and the downloaded filename.
Does generated code include JSON parsing methods?
Dart includes fromJson and toJson. Rust has Serde derives and C# has System.Text.Json key attributes. PHP and Scala outputs are declarations only, so your application supplies hydration or a codec.
Found a shape we should handle?
Generate a JSON to model draft
Run JSON to model with several realistic records, inspect the nullable fields and names, then download the current language's source file for project-side testing.