Skip to content
DocumentationStart buildingSign in
Generally available

Rerank

Put the right result first.

Open sourceOn-chain settlement

Cross-encoder reranking that reorders candidate documents by true relevance to the query. Drop it in after Vector or any search step to sharpen retrieval before the model spends a token on the wrong context.

Cross-encoder relevance scoring
Sharpens Vector & RAG retrieval
One API, open rerank models
Token-metered billing

The AI cloud — one API for every capability. Open models. On-chain.