Developers · GraphQL

GraphQL Importing/Exporting Overview

Last updated 1 November 2026

Export records and users to CSV, import them in bulk, and the CLI commands that do the same for a whole instance.

The import and export operations move data in bulk. Export records or users to CSV files in a zip, or load many records or users into an instance in one call. Use them for backups, for moving data between instances, and for loading data from another system.

The operations

Operation	Kind	What it does
`data_export`	Mutation	Starts a background export of records, users or both. Returns the export with its status
`data_exports`	Query	Lists exports, most recently updated first. A finished export holds the URL of its zip
`data_export_delete`	Mutation	Permanently deletes an export and its file
`import_models`	Mutation	Imports records in bulk. Each is updated if it matches, or created
`import_users`	Mutation	Imports users in bulk, the same way

How an export works

Start itdata_export selects records and users the same way the records and users queries do.
Wait for itThe export runs in a background job. Its status is pending, then done or failed.
Download itdata_exports returns the url of the zip once status is done.

An export is a background job, not an instant answer.

One export holds at most 5,000,000 records and 5,000,000 users. With split_files: true, the default, each table is split into files of at most 10,000 rows. Pass at least an empty object, such as records: {}, for that kind of data to be included.

Example: list the exports

Query

{
  data_exports(per_page: 3) {
    total_entries
    results { id status created_at url }
  }
}

Response on the docs instance

{
  "data": {
    "data_exports": {
      "total_entries": 0,
      "total_pages": 1,
      "results": []
    }
  }
}

This ran read-only on the docs.insites.io instance on 1 November 2026, which has no exports.

The CLI does the same for a whole instance

insites-cli data export staging -p data.zip

data export, data import and data clean work on all the data on an instance. There is no per-table option. data import reads a JSON file, or a zip archive with --zip.

Common mistakes

Reading url straight after data_export. It is empty until the status is done.
Leaving out records or users and getting an empty export.
Running insites-cli data clean to clear one table. It removes all data on the instance and cannot be undone.
Pairing the ids import_users returns with the people you sent by position. They can come back in a different order; match on external_ids or email.