Leto Schema

An implementation of GraphQL's type system in Dart. Supports any platform where Dart runs. The decisions made in the design of this library were done to make the experience as similar to the JavaScript reference implementation as possible, and to also correctly implement the official specification.

Contains functionality to build all GraphQL types:

  • String
  • Int
  • Float
  • Boolean
  • GraphQLObjectType
  • GraphQLUnionType
  • GraphQLEnumType
  • GraphQLInputObjectType
  • GraphQLListType
  • GraphQLNonNullType

Of course, for a full description of GraphQL's type system, see the official specification:

Mostly analogous to graphql-js; many names are verbatim:


It's easy to define a schema with the helper functions:

final todoSchema = GraphQLSchema(
  query: objectType(
    fields: [
      field('text', graphQLString.nonNull()),
      field('created_at', graphQLDate),

All GraphQL types are generic, in order to leverage Dart's strong typing support.

Table of Contents

GraphQL Schema

Each GraphQLSchema requires a GraphQLObjectType as the root query type and optional GraphQLObjectTypes for the mutation and subscription roots.

You can provide a list of directive definitions (GraphQLDirectives) that can be used within the schema or by documents executed by the schema. The directives will be extended if any of the types or fields have a ToDirectiveValue attachment.

A GraphQLSchema has an optional description String that can be used as documentation and a SerdeCtx for deserialization of input types. The astNode (SchemaDefinitionNode) will be set when the schema is created with buildSchema.

To validate the schema definition, following the specification, you can use the validateSchema(GraphQLSchema) function which returns the List<GraphQLError> found during validation.


Each field in an object type can provide a resolve callback to return the value when a GraphQL operation is executed over the schema.

final nameField = field(
  resolve: (Object parentObject, Ctx ctx) => 'Example Name',
  // or pass the subscribe parameter if it is a subscription.
  // The return type should be a Stream

For a more thorough discussion about resolvers please see the resolvers section in the main Documentation.

GraphQL Types

For a more thorough discussion about all the GraphQL types and their usage with Leto, please see the GraphQL Schema Types Documentation.

Scalar types

All of the GraphQL scalar types (GraphQLScalarType) are built in:

  • graphQLString
  • graphQLId
  • graphQLBoolean
  • graphQLInt
  • graphQLFloat

As well other additional types, provided for types in Dart's standard library:

Additional scalar types

  • graphQLDate
  • graphQLBigInt
  • graphQLTimestamp
  • graphQLUri

Composed Types

  • GraphQLObjectType
  • GraphQLUnionType
  • GraphQLEnumType
  • GraphQLInputObjectType
  • GraphQLListType
  • GraphQLNonNullType

Helpers and Extensions

These are helpers to create a GraphQL types within Dart. Most of them can be found in the /lib/src/gen.dart file.

  • objectType - Create a GraphQLObjectType with fields
  • field - Create a GraphQLField with a type/argument/resolver
  • inputObjectType - Creates a GraphQLInputObjectType
  • inputField - Creates a field for a GraphQLInputObjectType
  • listOf - Create a GraphQLListType with the provided innerType

Methods on GraphQLType

  • list - Create a GraphQLListType from the type
  • nonNull - Create a GraphQLNonNullType from the type
  • nullable - Returns the inner type of GraphQLNonNullType or itself it it is nullable
  • field - Create a GraphQLField (extension method)
  • inputField - Create a field for a GraphQLInputObjectType (extension method)

The extensions on GraphQLType with the field and inputField methods are recommended over the global function to preserve the GraphQLType's generic type information.

Serialization and SerdeCtx

GraphQL types can serialize and deserialize input data. The exact implementation of this depends on the type.

final serdeCtx = SerdeCtx();
final String iso8601String = graphQLDate.serialize(;
final DateTime date = graphQLDate.deserialize(serdeCtx, iso8601String);

Values passed through deserialize and serialize should round-trip.


A Serialization and Deserialization Context (SerdeCtx) allows you to create types from serialized values. It registers Serializers for any type that can be used with generics.


GraphQL types can validate input data. If the validation is successful, the GraphQLType.deserialize method should return an instance of the Dart type.

final validation = myType.validate('key', {...});

if (validation.successful) {
} else {

Non-Nullable Types

You can easily make a type non-nullable by calling its nonNull method.

List Types

Support for list types is also included. Use the list method from GraphQLType or the listOf helper function for convenience.

/// A non-nullable list of non-nullable integers

Input values and parameters

Take the following GraphQL query:

   anime {
     characters(title: "Hunter x Hunter") {

And subsequently, its schema:

type AnimeQuery {
  characters($title: String!): [Character!]

type Character {
  name: String
  age: Int

The field characters accepts a parameter, title. To reproduce this in package:leto_schema, use GraphQLFieldInput:

final GraphQLObjectType queryType = objectType(
  fields: [
      inputs: [
        GraphQLFieldInput('title', graphQLString.nonNull()),

final GraphQLObjectType characterType = objectType(
  fields: [
    field('name', graphQLString),
    field('age', graphQLInt),

In the majority of cases where you use GraphQL, you will be delegate the actual fetching of data to a database object, or some asynchronous resolver function.

package:leto_schema includes this functionality in the resolve parameter, which is a function that receives the parent object and a Ctx with a Map<String, dynamic> of input arguments.

A hypothetical example of the above might be:

final field = field(
  resolve: (_, Ctx ctx) async {
    final Stream<String> stream = await myDatabase.findCharacters(ctx.args['title']);
    return stream;

Schema and Document Validation Rules

GraphQL schemas and documents can be validated for potential errors, misconfigurations, bad practices or perhaps restrictions, such as restricting the complexity (how nested and how many fields) of a query. We perform all the document and schema validations in the specification. Most of the code was ported from graphql-js.

You can find the implementation for all the rules in the lib/src/validate/rules directory.

We also provide a QueryComplexity validation rule.

You can use two functions to validate GraphQL documents, both return a List<GraphQLError>:

validateDocument(GraphQLSchema, gql.DocumentNode)

You can specify multiple validation rules. The default is specifiedRules.

validateSDL(gql.DocumentNode, GraphQLSchema?)

If a GraphQLSchema is passed, it is assumed that the document SDL is an extension over the given schema.

You can specify multiple validation rules. The default is specifiedSDLRules(../leto_schema/lib/src/validate/validate.dart).

Custom Validations

You can provide custom validations, they are a function that receives a ValidationCtx and returns a visitor that reports errors thought the context.

For example, the following validation rule checks that argument names are unique:

const _uniqueArgumentNamesSpec = ErrorSpec(
  spec: '',
  code: 'uniqueArgumentNames',

/// Unique argument names
/// A GraphQL field or directive is only valid if all supplied arguments are
/// uniquely named.
/// See
Visitor uniqueArgumentNamesRule(
  SDLValidationCtx context, // ASTValidationContext,
) {
  final visitor = TypedVisitor();

  VisitBehavior? checkArgUniqueness(List<ArgumentNode> argumentNodes) {
    final seenArgs = argumentNodes.groupListsBy((arg) =>;

    for (final entry in seenArgs.entries) {
      if (entry.value.length > 1) {
            'There can be only one argument named "${entry.key}".',
            locations: List.of(entry.value
                .map((node) =>!.start)
                .map((e) => GraphQLErrorLocation.fromSourceLocation(e))),
            extensions: _uniqueArgumentNamesSpec.extensions(),

  visitor.add<FieldNode>((node) => checkArgUniqueness(node.arguments));
  visitor.add<DirectiveNode>((node) => checkArgUniqueness(node.arguments));
  return visitor;

GraphQLException and GraphQLError

A GraphQLException is a list of GraphQLErrors. You can view the GraphQLError definition in the following code section:

/// An error that may occur during the execution of a GraphQL query.
/// This will almost always be passed to a [GraphQLException].
class GraphQLError implements Exception, GraphQLException {
  /// The reason execution was halted, whether it is a syntax error,
  /// or a runtime error, or some other exception.
  final String message;

  /// An optional list of locations within the source text where
  /// this error occurred.
  /// Smart tools can use this information to show end users exactly
  /// which part of the errant query
  /// triggered an error.
  final List<GraphQLErrorLocation> locations;

  /// List of field names with aliased names or 0ÔÇÉindexed integers for list
  final List<Object>? path;

  /// The stack trace of the [sourceError]
  final StackTrace? stackTrace;

  /// An optional error Object to pass more information of the
  /// source of the problem for logging or other purposes
  final Object? sourceError;

  /// Extensions return to the client
  /// This could be used to send more
  /// information about the error, such as a specific error code.
  final Map<String, Object?>? extensions;

  /// An error that may occur during the execution of a GraphQL query.
  /// This will almost always be passed to a [GraphQLException].
    this.message, {
    this.locations = const [],
    StackTrace? stackTrace,
  }) : stackTrace = stackTrace ?? StackTrace.current;

Ctx and ScopedMap

You can view a more thorough explanation in the main README.


You can view an usage example in the main README.


Most GraphQL utilities can be found in the utilities folder in package:leto_schema.


Create a GraphQLSchema from a GraphQL Schema Definition (SDL) document String.


Transform a GraphQLSchema into a String in the GraphQL Schema Definition Language (SDL).


Experimental. Extend a GraphQLSchema with an SDL document. This will return an extended GraphQLSchema with the additional types, fields, inputs and directives provided in the document.


Create an introspection document query for retrieving Schema information from a GraphQL server.


Experimental. Merge multiple GraphQLSchema. The output GraphQLSchema contains all the query, mutations and subscription fields from the input schemas. Nested objects are also merged.


Experimental. Build a GraphQLSchema from a JSON value, will add query, mutation, subscription and custom events on top of the provided JSON value. Will try to infer the types from the JSON structure.