Query HubSpot data using GraphQL (BETA)

Last updated:

This functionality is in beta and requires your account to be ungated. You can request access by submitting this form. By participating in this beta you agree to HubSpot's Developer Terms and Developer Beta Terms. Please note that this functionality is currently under development and is subject to change based on testing and feedback.

If you encounter any issues during the beta, you can get help in the data-driven-content channel on the HubSpot Developer Slack community.

  • CMS Hub
    • Professional or Enterprise

GraphQL is a data query language that provides a unified way to access your HubSpot CRM and HubDB data to create personalized and data-driven experiences for your website visitors.

Below, learn more about the benefits of using GraphQL, how to structure and test GraphQL queries, and what limits are placed on the complexity of your queries. Once you construct a query that retrieves the data you need, you can use it directly in your website pages by binding it to a theme or module.

If you want to see an example of how GraphQL queries work in practice, check out this tutorial.

Advantages of using GraphQL

The GraphQL API provides a single, unified way to fetch data from different data sources to power your website pages, instead of having to process and aggregate data from disparate REST API endpoints.

You can specify the data within GraphQL queries, which allows you to decouple how you're fetching data from the presentational layer of your template and module code. You can then include queries as part of your theme, or connect them directly to your modules, templates, and pages.

GraphQL schema and queries

In HubSpot, the GraphQL schema describes the data sources you're fetching from and the underlying data types available to your HubSpot account. This schema is auto-generated and will automatically update if you add custom properties, create custom objects, or associate object types.

You can construct GraphQL queries to fetch data by traversing the schema from your HubSpot account. This approach allows you to specify the data you need for a given page on your website.

Each GraphQL query consists of a tree of nested fields: 

query MyQuery { # Root query field CRM { # Child fields contact_collection { items { # Properties email firstname lastname } } } }
  • Root query field: the top-level field of your query that represents your data source. Currently, the only supported data source is the HubSpot CRM.
  • Child fields: based on the data source in your root query field, you can then specify the child fields you want to fetch, along with any associated properties.

Follow the instructions in the sections below to fetch data from the HubSpot CRM or a HubDB table. You can test and run your query interactively using GraphiQL, which provides autocompletion and allows you to browse all available fields you can include in your query.

Query data from your CRM

To query data from the HubSpot CRM, include CRM as a root query field, then enter a data type as a child field. You can include any of the following data types from the CRM:

  • Standard and custom objects
  • Engagements, including calls, tasks, meetings, and notes
  • Products and line items
  • Quotes and subscriptions
  • Marketing events

You can retrieve a single instance of a data type, or enter a data type followed by _collection to retrieve a list of objects of that type. You can provide arguments to a query field to fetch a single object or filter a collection of objects.

  • Fetch a single instance: if you're fetching a single instance like an individual contact record, specify one of the object's properties and its associated value by including the property's name as the uniqueIdentifier, and the property's value as the uniqueIdentifierValue. For example, the following query would fetch a contact with an ID of 753:
query filteredContact { CRM { contact(uniqueIdentifier: "id", uniqueIdentifierValue: "756") { email } } }
  • Fetch and filter a collection: to fetch a collection of instances, add items as a child field. To filter the objects in the response, include the filter argument, then provide a map of filter operators to their associated values. A filter operator consists of the property you want to filter by, followed by a suffix. For example, the query below would fetch and filter a collection of contacts with hubspot.com email addresses:
query filteredContactCollection { CRM { contact_collection(filter: {email__contains: "hubspot.com"}) { items { email } } } }

Please note: if you're using the __in operator to filter on a multiple checkbox property, the data in the response will include instances that partially match one or more of the values in your filter argument.

For example, if your query field was contact_collection(filter: {hs_buying_role__in: ["END_USER", "INFLUENCER"]}), then the response would include contacts who had End User selected as one of their buying roles, but didn't necessarily have Influencer as one of their buying roles.

After you've added a data type to your query, you can specify the properties, associations, or metadata you need by adding child fields:

  • Add properties to fetch: enter the internal name of a property to include it in the response of your query. Both default and custom properties are supported.
  • Include associations to other data types: you can also add a collection of instances of a data type associated with the parent data type, such as the companies associated with a contact. To add data from an associated data type:
    • Include associations as a child field of the original data type you entered.
    • As a child field of associations, enter the association you want to include by following the naming scheme of {ASSOCIATED_DATA_TYPE}_collection__{LABEL_SUFFIX}.
      • The associated data type can be a built-in association with a standard object, or an association with a custom object. Any associations to custom objects will be prefixed by p_.
      • For primary associations, the label suffix is primary. If you created a custom association label, the label suffix will be lower-cased, with each word in the label separated by an underscore. For example, if your custom association label is My custom association label, the label suffix would be my_custom_association_label.
      • For example, if you had a custom object named House associated with the contact object, the association you'd include in your query would be p_house_collection__house_to_contact.
    • Add items as a child field of your association, then include any properties you need from each item in the collection.
query contactsWithAssociatedCompany { CRM { contact_collection { items { firstname lastname associations { company_collection__primary { items { name address } } } } } } }

Query data from HubDB

Please note: if HUBDB doesn't appear as a data source or any published tables are missing when you're using GraphiQL, try logging out of your HubSpot account and logging back in.

To query data from HubDB, include HUBDB as a root query field, then enter the name of a table to retrieve a single row from that table, or enter a name followed by _collection to retrieve multiple rows from that table. You can provide arguments to a query field to retrieve a single row or filter multiple rows from the table:

  • Retrieve a single row: if you're retrieving a single row from a table, specify one of the table's columns as the uniqueIdentifier then provide the associated value of the row you want to retrieve as the uniqueIdentifierValue. For example, the following query would retrieve a row from an executives table with a role of CEO:
query filteredExecutive { HUBDB { executives(uniqueIdentifier: "role", uniqueIdentifierValue: "CEO") { name email description } } }

Please note: if the column you provide as the uniqueIdentifier has non-unique values, your query will only return the first row, ordered by row ID.

  • Retrieve and filter multiple rows: to fetch multiple rows from a table, add items as a child field. To filter the rows in the response, include the filter argument to your query field, then provide a map of filter operators to their associated values. A filter operator consists of the property you want to filter by, followed by a suffix. For example, the query below would retrieve and filter rows from the executives table whose role contains Officer:
query csuiteExecutives { HUBDB { executives_collection(filter: {role__contains: "Officer"}) { items { email } } } }

After you've added a data type to your query, you can specify the columns or metadata you need by adding child fields:

  • Add columns to retrieve: enter the name of a column to include it in the response to your query. If one of your columns is a foreign ID and you want to include columns from the associated table:
    • Include associations as a child field of items.
    • As a child field under associations, enter the name of the foreign table with the _collection suffix, followed by two underscores and the name of the foreign ID column. For example, if the foreign ID column is recent_posts, and the name of the foreign table you want to reference is blog_posts, then the resulting field to include in your query would be blog_posts_collection__recent_posts.
    • Add items as a child field of the foreign table field you just entered, then include any columns you need from the foreign table. For example, to include additional columns from the executives table from the csuiteExecutives query above, along with the title and link from a blog_posts foreign table, you'd update the query to the following:
query csuiteExecutives { HUBDB { executives_collection(filter: {role__contains: "Officer"}) { items { associations { blog_posts_collection__recent_posts { items { title link } } } email name description role } } } }
  • Include metadata fields: internal properties of the table, such as the ID or the timestamp of the last update to a row, are prefixed with hs_.

Retrieving data from child tables

If you've configured dynamic pages to be created from your HubDB tables, you can include data from child tables in your GraphQL query.

  • Add hs_child_table_collection as a child field of the items of the parent table collection.
  • For each child table that you want to include, include a nested child fragment, which consists of the spread operator (i.e., ...) combined with a type condition, which includes the on keyword and the name of the child table collection. For example, if you wanted to update the csuiteExecutives query from above to include columns from the team and recent_posts child tables:
query csuiteExecutives { HUBDB { executives_collection(filter: {role__contains: "Officer"}) { items { email name description role hs_child_table_collection { ... on teams_collection { name email summary } ... on recent_posts_collection { title summary link } } } } } }

Retrieving unpublished data from a table

You can include unpublished data from a table by providing {draft: true} as a query argument to the name of the table that you're retrieving. Any new but unpublished columns you've added will not be included. 

query csuiteExecutives { HUBDB { executives_collection(draft: true, filter: {role__contains: "Officer"}) { items { email name description role hs_id hs_child_table_id } } } }

Test and run queries interactively using GraphiQL

You can test queries by using GraphiQL, which is an open-source tool that allows you to run queries interactively and explore which fields are available.

  • In your HubSpot account, navigate to the GraphiQL tool.
  • In the left pane, create and edit your query by specifying the objects and their associated fields you'll need for your website pages.


  • To view all available data types, properties, and filters, click Explorer at the top of the page to toggle the Explorer view, which will list all available data types and their associated properties. 
    • Click a data type to automatically add it to your query and view all its associated properties.
    • Select the checkbox next to a property to add it to your query.
    • If you're writing a query and you're entering an argument for one of your fields, the Explorer pane will auto-populate a list of available filters.
  • When you're ready to test a query, click the play icon at the top of the page. The query's response will appear in the right pane.



Query arguments and filters

You can provide arguments to a collection field in your query to filter, order, or paginate the objects returned in the response. The table below lists the supported arguments:

Use this table to describe parameters / fields
Input type

A set of filter options to apply on the collection items


The maximum number of items to fetch.

If your query's data source is the CRM, the default limit is 10 items, and the maximum limit is 500.

If you're using HUBDB as a data source, the default limit is 1000 rows, with no maximum limit.

Enum array

The ordering to apply on the fetched items. The default ordering is ascending by the item's id.


The index from where to start fetching items. The default is 0. When paginating results, the next offset is returned in the response.

Create a filter by combining a HubSpot property with a suffix that corresponds to a logical or mathematical operator, separated by two underscores (e.g., email__contains or name__eq). The supported operators are shown in the table below:

Filter Postfix Field types Example
equal __eq String, Number, Date, DateTime email__eq: “user@domain.com”
not equal __neq String, Number, Date, DateTime firstname__neq: “Bob”
less than __lt Number, DateTime price__lt: 45
less than or equal __lte Number, DateTime started__lte: 1633606374
greater than __gt Number, DateTime salary__gt: 60000
greater than or equal __gte Number, DateTime birthdate__gte: 1633606374
contains __contains String name__contains: “Inc.”
does not contain __not_contains String address_not_contains: “Miami”
in given list __in Enumeration, Number hs_buying_role__in: [“END_USER”, "INFLUENCER"]
not in given list __not_in Enumeration, Number status__not_in: [“REJECTED”, "PENDING"]
has value / does not have value __null Any type email__null: false

Query complexity and account limits

To ensure optimal performance for querying your data, HubSpot enforces a limit on the aggregate complexity of your GraphQL queries, based on factors such as the total number of objects and their associated properties in your query. Likewise, since HubSpot needs to make a new internal API request for each top-level object, complex queries with nested associations incur an additional cost to execute. 

The following factors are assigned point values and used to calculate a resulting complexity score:

  • Internal API request: 300 points
  • Object retrieved: 30 points
  • Requested property with a value: 3 points
  • Requested property without a value: 1 point

HubSpot will multiply the number of occurrences of each factor listed above by its corresponding point value, then sum each of these subtotals to arrive at a final complexity score.

For example, the following query retrieves contacts, their associated companies, and the tickets associated with those companies:

query myQuery { CRM { contact_collection { items { hs_object_id firstname lastname email company associations { company_collection__primary { items { hs_object_id name domain country annualrevenue phone associations { ticket_collection__primary { items { hs_object_id createdate content created_by hs_pipeline hs_pipeline_stage closed_date } } } } } } } } } }

Assuming that the above query retrieves 10 contacts, with 6 companies per contact, and 15 tickets per company, the complexity score would be calculated as follows:

  • Points used for contact_collection:
    • One internal API request is made (300 points).
    • 10 contacts are retrieved (10 * 30 = 300 points).
    • For each contact, five properties with values are retrieved (10 * 5 * 3 =150 points).
    • Total points: 750
  • Points used for company_collection__primary:
    • One internal API is made per contact retrieved, totaling 10 requests (10 * 300 = 3,000 points).
    • Six companies are retrieved (6 * 30 = 180 points).
    • Five properties are retrieved with a value per company (6 * 5 * 3 = 90 points).
    • One property is returned with no value per company (6 * 1 = 6 points).
    • Total points: 3,276
  • Points used for ticket_collection__primary:
    • One internal API request is made per company per contact retrieved (10 * 6 * 300 = 18,000 points).
    • 15 tickets are returned (15 * 30 = 450 points).
    • Seven properties are returned with a value per ticket (15 * 7 * 3 = 315 points).
    • Total points: 18,765

Adding the points from the three constituent parts of the query, the final complexity score for the query totals to 22,791 points. 

In addition to the complexity limits on individual queries, HubSpot also enforces an account-wide limit for the complexity of the queries you include when using the external GraphQL API endpoint:

  • Each individual query has a maximum of 30,000 points available. If this limit is reached while the query is running, all further execution is blocked and all object instances fetched up to that point are returned.
  • Each HubSpot account is allocated up to 500,000 points per rolling minute when using the GraphQL API endpoint. Accumulated points from the previous minute are not transferred to the next. If your account has used all available points per rolling minute, all further requests will receive an HTTP status code of 429 Too Many Requests.
  • The normal burst limits detailed in HubSpot's API usage guidelines are not applicable to querying data using GraphQL.

When HubSpot retrieves the data specified in your GraphQL query, the response contains details on the complexity score for your query within the extensions field.

  • The complexity score for the query you provided in your request will be broken down within the query_complexity field.
  • If you're using the external GraphQL API endpoint, the extensions field will also include account limit information within rate_limit_info.
"extensions": { "rate_limit_info": { "interval_in_seconds": 60, "max_points": 500000, "remaining_points": 477209, } "query_complexity": { "max_points": 30000, "used_points": 22791, "points_for_internal_api_requests": { "count": 71, "weight": 300, "used_points": 21300 }, "points_for_objects_retrieved": { "count": 31, "weight": 30, "used_points": 930 }, "points_for_properties_with_value": { "count": 185, "weight": 3, "used_points": 555 }, "points_for_properties_without_value": { "count": 6, "weight": 1, "used_points": 6 } } }

Use a GraphQL query in an API request

To use a GraphQL query in your integration, a Jamstack site, or in a custom code action in the workflows tool, you can include it in an authenticated POST request to the /collector/graphql endpoint.

In the request body, include the operationNamequery, and any variables you want to pass to the query.

Use this table to describe parameters / fields

A label for the query included in the request


The GraphQL query to execute

JSON object

A JSON object containing any variables you want to pass to the query

For example, if you created a House custom object, and you wanted to query for a collection of House records in Miami, the request body would contain:

// POST request body to https://api.hubapi.com/collector/graphql?hapikey=xxxx-xxxxx-xxxxx-xxxxx-xxxxx { "operationName": "houses" "query": "query houses ($city: String) { CRM {house_collection(filter: {city__eq: $city}) { items { asking_price bedrooms bathrooms city}}}}", "variables": {"city": "Miami"} }

Example return payload:

{ "data": { "CRM": { "house_collection": { "items": [ { "asking_price": 800000, "bedrooms": 3, "bathrooms": 2, "city": "Miami" } { "asking_price": 940000, "bedrooms": 2, "bathrooms": 2, "city": "Miami" } ] } } } }

Further reading

Once you're familiar with how to create a GraphQL query, check out this guide to using a query on your website pages.

Was this article helpful? *
This form is used for documentation feedback only. Learn how to get help with HubSpot...