Codable, JSON, and API models | Swift

Apps often exchange data as JSON. Swift’s Codable protocols let you convert between JSON and Swift types.

A Codable model

struct Article: Codable {
    let id: String
    let title: String
    let authorName: String
}

If JSON keys match property names and property types are also Codable, Swift can synthesize encoding and decoding code.

Decoding JSON

let json = """
{
    "id": "a-1",
    "title": "Swift optionals",
    "authorName": "Maya"
}
"""

let data = Data(json.utf8)

let article = try JSONDecoder().decode(Article.self, from: data)

The result is a real Article, not a loose dictionary. After decoding succeeds, the compiler knows article.id is a String.

Mapping API names

APIs often use snake_case keys:

{
  "author_name": "Maya"
}

You can configure a decoder:

let decoder = JSONDecoder()
decoder.keyDecodingStrategy = .convertFromSnakeCase

Or define custom coding keys:

struct Article: Codable {
    let authorName: String

    enum CodingKeys: String, CodingKey {
        case authorName = "author_name"
    }
}

API models vs domain models

An API shape is not always the best app shape. It is common to decode a transport model, then convert it:

struct ArticleResponse: Codable {
    let id: String
    let title: String?
}

struct Article {
    let id: String
    let title: String
}

This lets the app enforce stronger rules internally, such as “an article title is always present.”

What to carry forward

Codable combines Encodable and Decodable
JSON can be decoded into typed Swift models
decoding can fail, so it uses try
API naming can be mapped with decoder strategies or coding keys
separate API models from app models when external data is messy

Next, you will run simple Swift programs from the command line.