Embed 1-100 texts into vectors using Cloudflare Workers AI BGE models (`bge-base` 768d default, `bge-large` 1024d, or multilingual `bge-m3` 1024d). Returns one vector per input text. Priced per text.
Reorder up to 50 candidate documents by relevance to a query using Cloudflare Workers AI `@cf/baai/bge-reranker-base`. Returns documents with relevance scores, highest first. Pair with /embed to build a retrieve-then-rerank RAG pipeline.