Shaped

Overview

Shaped is an AI-native personalization and hybrid search platform that provides the functionality of a vector database plus additional ranking, experimentation, and business-objective control layers out of the box.

Features

• Hybrid search: keyword + vector retrieval fused together for relevance
• Ranking & personalization: embeddings, multi-stage ranking (two-tower, ANN, transformers), and real-time session adaptation
• Value Modeling: blend conversions, engagement, diversity, freshness, and fairness at inference time without retraining
• Warehouse-native: plug directly into Snowflake, BigQuery, Redshift with SQL transforms for transparent feature engineering
• Real-time APIs: personalized feeds, recommendations, and search endpoints ready for production
• Cold-start handling: uses metadata, hybrid retrieval, and transfer learning to serve relevant results from day one
• Real-time updates: supports ingesting and re-ranking in real time, adapting within-session

Pricing

Contact-based pricing; no public self-serve tier listed.