Retrieve personalized recommendations for your users in real time.

This endpoint lets you retrieve a given number of personalized recommendations for a specific user.
The recommendation endpoint takes two required parameters:

  • the number of results to retrieve
  • the unique user identifier for which to retrieve these results. This is either the persistent user ID of the user, or their semi-persistent anonymous ID (see our guide on identifying users for details on how to find this id).

The number of results is guaranteed up to 100 or the number of eligible items available in your dataset, depending on whichever is smaller.

You can optionally provide additional information about the current user context. Supported parameters for the context attribute are:

  • The itemId of the item that the user is currently interacting with.
  • The current user location given as latitude and longitude.

If successful, you will receive a JSON response containing a list of items, each being given as an object with an itemId and probability. Note that the probability indicates the algorithm’s confidence in the recommendation but does not necessarily equate to the probability that the user will engage with the recommended item.

The response also includes a correlationId. This is a unique identifier for the recommendation request, and should be sent as part of your analytics in the _flike_correlation_id property when users engage with a recommended item (e.g. click on it or otherwise interact with it).

Click Try It! to start a request and see the response here!