GraphQL API

GraphQL Introduction

GraphQL is a query API language that executes queries by using a type system based on defined input data. GraphQL provides more specific inquiries with faster results and populate multiple inputs into one query.

Note: GraphQL endpoint permissions mirror permissions for the REST API. A user is only granted access to view grades based on that user’s permissions. For instance, a student cannot view grades for another student, but an instructor can view grades for any student in a course.

Learn more about GraphQL.

Using GraphQL

Canvas has included the tool GraphiQL (https://github.com/graphql/graphiql), an in-browser graphical interface for interacting with GraphQL endpoints.

The GraphiQL interface can be viewed by adding /graphiql to the end of your Canvas production URL (e.g. your-institution.instructure.com/graphiql).

The /graphiql access can also be added to a test or beta environment URL. Requests from the selected environment will always return that environment’s data.

The Explorer sidebar displays all available queries and mutations. Any selected items display in the GraphiQL window. Once a query or mutation is selected, any values displayed in purple text identify the value as an input argument.

REST vs GraphQL

The Canvas REST API will continue to be available. New Canvas features will be developed primarily in GraphQL and may not be back-ported to the REST API.

Fields are being added to the GraphQL API on an as-needed basis. The GraphQL API does not include everything that is currently in the REST API. Feel free to submit pull requests on github to add additional features or talk about it in the #canvas-lms channel on freenode.

GraphQL Endpoint

POST /api/graphql

All GraphQL queries are posted to this endpoint.

Request Parameters

Parameter Type Description
query string the GraphQL query to execute
variables Hash variable values as required by the supplied query

Example Request:

curl https://<canvas>/api/graphql \
  -H 'Authorization: Bearer <ACCESS_TOKEN>' \
  -d query='query courseInfo($courseId: ID!) {
       course(id: $courseId) {
        id
        _id
        name
       }
     }' \
  -d variables[courseId]=1

Example Response

{
  "data": {
    "course": {
      "id": "Q291cnNlLTE=",
      "_id": "1",
      "name": "Mr. Ratburn's Class"
    }
  }
}

GraphQL in Canvas

id vs _id and the node field

The Canvas LMS GraphQL API follows the Relay Object Identification spec. Querying for an object's id will return a global identifier instead of the numeric ids that are used in the REST API. The traditional ids can be queried by requesting the _id field.

Most objects can be fetched by passing their graphql id to the node field:

{
  node(id: "Q291cnNlLTE=") {
    ... on Course {
      _id  #  traditional ids (e.g. "1")
      name
      term { name }
    }
  }
}

A legacyNode field is also available to fetch objects via the REST-style ids:

{
  # object type must be specified when using legacyNode
  legacyNode(type: Course, _id: "1") {
    ... on Course {
      _id
      name
    }
  }
}

For commonly accessed object types, type-specific fields are provided:

{
  # NOTE: id arguments will always take either graphql or rest-style ids
  c1: course(id: "1") {
    _id
    name
  }
  c2: course(id: "Q291cnNlLTE=") {
    _id
    name
  }
}

Pagination

Canvas follows the Relay Connection Spec for paginating collections. Request reasonable page sizes to avoid being limited.

{
  course(id: "1") {
    assignmentsConnection(
      first: 10,      # page size
      after: "XYZ"    # `endCursor` from previous page
    ) {
      nodes {
        id
        name
      }
      pageInfo {
        endCursor     # this is your `after` value for the next request
        hasNextPage
      }
    }
  }
}