Generate personalized Polymarket market recommendations for a wallet address.
This endpoint uses a two-stage recommendation architecture:
-
Candidate Generation: Filters, boosts, and sorts markets to generate candidate recommendations
-
Reranking: Personalizes and reorders candidates based on user behavior
Candidate Filtering:
Use filters.candidate_filters for advanced filtering and relevance boosting.
Filter Types:
-
term: Exact match for single value (e.g.,{"filter": "term", "field": "active", "value": true}) -
terms: Match any of multiple values (e.g.,{"filter": "terms", "field": "ai_labels_high", "value": ["mbd2:t_sports"]}) -
numeric: Numeric comparison (e.g.,{"filter": "numeric", "field": "volume_1yr", "value": 100, "operator": ">="}) -
date: Date range filter (e.g.,{"filter": "date", "field": "end_date", "value": {"date_from": "", "date_to": "2026-01-01"}}) -
match: Text search (e.g.,{"filter": "match", "field": "question", "value": ["bitcoin"]})
Scoring Methods:
-
none: No reranking, returns candidates in original order -
hrnn_v0: Personalized reranking based on user trading history and behavior