FEATURES API

Turn wallet behavior into ranking features.

Compute wallet-item signals your app or models can use: semantic similarity, topic similarity, preference matches, labels, tags, and market-specific activity features.

Dense vectors

Semantic similarity between wallets and items

Topics & tags

Label and preference matching per wallet

ML-ready

Normalized features for any ranking model

01

Send a user and candidates

Pass a wallet profile and the items you are considering — markets, tokens, traders, or posts.

02

Embed computes the signals

Semantic similarity, topic overlap, preference matches, and market activity features, normalized for ranking.

03

Use them anywhere

Feed the features into Rank, your internal ML, lifecycle models, or analytics.

Embeddings are the feature, not the product.

Pretrained onchain embeddings power the Features API — plug them into ranking, fraud, lifecycle, and growth models.