Overview
This node integrates with the Recombee recommendation API to recommend users for a specific item. It is designed to handle multiple input items in batches, retry failed requests, and return detailed recommendation results. This node is useful when you want to identify which users are most likely to be interested in or related to a particular item, such as recommending customers who might like a product or content piece.
Practical examples:
- An e-commerce platform wants to find users who would be interested in a newly added product.
- A media streaming service recommends viewers who might enjoy a specific movie or show.
- A marketing team targets users for personalized campaigns based on an item of interest.
Properties
| Name | Meaning |
|---|---|
| Item ID | The unique identifier of the item to recommend users to. |
| Return Properties | Whether to include the properties/details of the recommended users in the output. |
| Count | The number of user recommendations to return per item. |
| Scenario | The recommendation scenario or context to use (e.g., "homepage", "relatedUsers"). |
| Cascade Create | Whether to create the item in Recombee if it does not already exist. |
| Batch Size | Number of requests to send concurrently in one batch. |
| Max Retries | Maximum number of retry attempts for failed requests. |
| Timeout (Ms) | Request timeout duration in milliseconds. |
Output
The node outputs an array of JSON objects, each corresponding to an input item processed. Each object contains:
success: Boolean indicating if the recommendation request succeeded.itemId: The item ID used for the recommendation.count: Number of recommended users requested.scenario: The scenario parameter used.returnProperties: Whether properties were returned.data: The actual recommendation data returned by Recombee, typically including user IDs and optionally their properties if requested.error(only if failed): Error message describing what went wrong.
The output does not include binary data.
Dependencies
- Requires an API key credential for authenticating with the Recombee API.
- Needs configuration of:
- Recombee database ID
- Recombee private token
- Recombee database region
- Optional API timeout setting (default 10 seconds)
- Uses the official Recombee API client library.
Troubleshooting
Common issues:
- Invalid or missing API credentials will cause authentication failures.
- Incorrect item IDs may result in errors or empty recommendations.
- Network timeouts or rate limits can cause request failures; increasing timeout or reducing batch size may help.
- If the item does not exist and cascade create is false, recommendations may fail.
Error messages:
- Errors from the Recombee API are caught and reported in the output's
errorfield. - Node operation errors will halt execution unless "Continue On Fail" is enabled.
- Retry logic attempts failed requests up to the configured max retries before throwing errors.
- Errors from the Recombee API are caught and reported in the output's