About the GraphQL Explorer
GraphiQL, also referred to in this documentation as the GraphQL Explorer, is a "graphical interactive in-browser GraphQL IDE."
Using the Altair GraphQL Client IDE
There are many open source GraphQL client IDEs. For example, you can use Altair to access GitHub's GraphQL API. To access the GraphQL API with Altair, download and install it from altair-graphql/altair. Then, follow the configuration steps below.
Configuring Altair
- Get an access token.
- Launch Altair.
- In the left sidebar, below the Altair logo, click Set Headers. A new window will open.
- In the "Header key" field, enter
Authorization
. - In the "Header value" field, enter
Bearer TOKEN
, replacingTOKEN
with your token from the first step. - Click Save in the bottom right corner of the window to save your authorization header.
- In the "GraphQL Endpoint" field, enter
http(s)://<em>HOSTNAME</em>/api/graphql
. - To load the GitHub GraphQL schema, download the public schema.
- In Altair, click on Docs on the top right, then the three dots and Load Schema...
- Select the file public schema that you downloaded in an earlier step.
Note: For more information about why POST
is the method, see "Forming calls with GraphQL."
You can test your access by querying yourself:
query {
viewer {
login
}
}
If everything worked correctly, this will display your login. You're all set to start making queries.
Accessing the sidebar docs
All types in a GraphQL schema include a description
field compiled into documentation. The collapsible Docs pane on the right side of the Explorer page allows you to browse documentation about the type system. The docs are automatically updated and will drop deprecated fields.
The Docs sidebar contains the same content that is automatically generated from the schema under "GitHub GraphQL API documentation," though it is formatted differently in places.
Using the variable pane
Some example calls include variables written like this:
query($number_of_repos:Int!){
viewer {
name
repositories(last: $number_of_repos) {
nodes {
name
}
}
}
}
variables {
"number_of_repos": 3
}
This is the correct format to submit the call using a POST
request in a curl
command (as long as you escape newlines).
If you want to run the call in the Explorer, enter the query
segment in the main pane and the variables in the Query Variables pane below it. Omit the word variables
from the Explorer:
{
"number_of_repos": 3
}
Requesting support
For questions, bug reports, and discussions about GitHub Apps, OAuth apps, and API development, explore the APIs and Integrations discussions on GitHub Community. The discussions are moderated and maintained by GitHub staff, but questions posted to the forum are not guaranteed to receive a reply from GitHub staff.
Consider reaching out to GitHub Support directly using the contact form for:
- guaranteed response from GitHub Enterprise Server staff
- support requests involving sensitive data or private concerns
- feature requests
- feedback about GitHub Enterprise Server products
Troubleshooting errors
Because GraphQL is introspective, the Explorer supports:
- Intelligent typeaheads aware of the current schema
- Validation error previews as you type
If you enter a query that is not well-formed or does not pass schema validation, a popup warns you of an error. If you run the query, the error returns in the response pane.
A GraphQL response contains several keys: a data
hash and an errors
array.
{
"data": null,
"errors": [
{
"message": "Objects must have selections (field 'nodes' returns Repository but has no selections)",
"locations": [
{
"line": 5,
"column": 8
}
]
}
]
}
It's possible you might run into an unexpected error that is not related to the schema. If this happens, the message will include a reference code you can use when reporting the issue:
{
"data": null,
"errors": [
{
"message": "Something went wrong while executing your query. This is most likely a GitHub bug. Please include \"7571:3FF6:552G94B:69F45B7:5913BBEQ\" when reporting this issue."
}
]
}
Note: GitHub recommends checking for errors before using data in a production environment. In GraphQL, failure is not total: portions of GraphQL queries may succeed while others fail.