Automatic Encoding/Decoding

The @json annotation will provide the decoder and encoder for that data class, so we are able to read from and write to Json.

data class Person(val name: String, val age: Int) {
  companion object // <- This is needed

You will be able to encode and decode using the following:

import arrow.core.*
import helios.*
import helios.core.*
import helios.typeclasses.*
val personJson = with(Person.encoder()) { Person("Abc", 10).encode() }

Building a Json

You can build your own Json object like this:

val jObject = JsObject(
"name" to JsString("Elia"),
"age" to JsNumber(23)

Custom Encoders

To create a custom Encoder, you need to inherit from the Encoder interface and implement the encode method.

val personCustomEncoder = object : Encoder<Person> {
  override fun Person.encode(): Json =
      "first_name" to JsString("John"),
      "age" to JsNumber(28)

val personCustomJson = with(personCustomEncoder) { Person("Abc", 10).encode() }

Custom Decoders

You can follow the same approach to create a custom Decoder:

import arrow.core.extensions.either.applicative.applicative
import helios.instances.decoder

val personCustomDecoder = object : Decoder<Person> {
  override fun decode(value: Json): Either<DecodingError, Person> =
      value["first_name"].fold({ Either.Left(KeyNotFound("first_name")) }, { it.decode(String.decoder()) }),
      value["age"].fold({ Either.Left(KeyNotFound("age")) }, { it.decode(Int.decoder()) })
    ) { tuple ->
      Person(tuple.a, tuple.b)


You can navigate Json using the Json.path DSL to select keys or traverse collections.

import helios.optics.*"name").string.modify(jObject, String::toUpperCase)

Note that the code generation will give you an accessor for each json field., String::toUpperCase)