JSON to Go Struct

Build Go struct definitions with proper JSON tags from any JSON sample. Handles nested objects via inline structs, arrays as slices, and ISO date strings as <code>time.Time</code>. Output is ready to paste into a Go source file — compiles without modification.

Why generate Go structs

Hand-typing a 30-field struct with correct tags is busywork. Tools like mholt/json-to-go solved this years ago for Go developers; this is the same idea, in your browser, with a few modern niceties (time.Time inference, omitempty toggling, named-vs-inline).

What you get

type User struct {
    ID        int    `json:"id"`
    Email     string `json:"email"`
    Name      string `json:"name"`
    Tags      []string `json:"tags,omitempty"`
    Address   struct {
        Street string `json:"street"`
        City   string `json:"city"`
        State  string `json:"state"`
        Zip    string `json:"zip"`
    } `json:"address"`
    CreatedAt time.Time `json:"created_at"`
}

Common workflows

Type an API client. Generate structs from the API’s sample response, drop into a types.go, use with json.Unmarshal or your client library.

Stub a service handler. Generate request and response structs from a contract example, write the handler against them.

Move from interface{} to typed. A Go service using map[string]interface{} can be partially typed by generating structs from a sample and refactoring incrementally.

Edge cases worth knowing

• Sums / discriminated unions — Go has no native enum, no native sum type. The tool produces a string field; you add a custom UnmarshalJSON if you need typed variants.
• Numbers without decimal — typed as int. If you receive 1.0 as a number, that becomes int; if your data later includes 1.5, the generated type won’t fit. Inspect a richer sample.
• Empty arrays — typed as []interface{} (we cannot infer element type from an empty array). Add a non-empty sample.

For runtime validation alongside the types, pair with a validation library (go-playground/validator).

Last updated: 2025-01-15