Setup
1
Choose item ID ownership
Decide whether your source catalog will provide stable item IDs or let NeuronSearchLab generate them.
2
Map source fields
Map source fields to
/v1/items payloads (name, description, metadata).3
Emit recommendation events
Emit user events to
/v1/events from all recommendation surfaces.Common workflows
Create catalog items
Submit user interaction events
Maintain ingestion quality
- Validate required fields before publishing payloads.
- Use Unix timestamps in integer seconds for event consistency.
- Monitor event throughput and lag in your own pipeline metrics.
Example app journey: Daily sync + real-time feedback
1
Sync catalog records
Nightly job creates new catalog items using
/v1/items and updates existing records with /v1/items/{id}.2
Request recommendations
App session requests recommendations from the Core API recommendation endpoint.
3
Forward interactions
Frontend interaction events are forwarded to backend.
4
Publish events
Backend publishes click/purchase events to
/v1/events in near real time.5
Review analytics
Team reviews Analytics to confirm engagement trends are improving.

