docs / embeddings

Text Embeddings

Turn text into vector embeddings.

base /embeddings/v12 endpoints

post/embeddings/v1/embed1 credit

text -> dense embedding vector(s). input = a string OR a list of strings (Qwen3-Embedding-4B, 2560-dim, multilingual). Returns embeddings[] + dim + usage.

Parameter	Required
input	required
model	optional

Try in playground →

post/embeddings/v1/rerank1 credit

rerank documents by relevance to a query (BGE-reranker-v2-m3 cross-encoder, multilingual). params: query (str) + documents (list of str). Returns ranked results[{index, relevance_score}] (most-relevant first).

Parameter	Required
query	required
documents	required
model	optional
top_n	optional

Try in playground →

Example request · embed

curl -X POST https://api.reefapi.com/embeddings/v1/embed \
  -H "x-api-key: $REEF_KEY" \
  -H "content-type: application/json" \
  -d '{"input":["best coffee shop reviews near me"]}'

Response shape

{
  "ok": true,
  "data": { /* the result */ },
  "meta": {
    "latency_ms": 240,
    "record_count": 12,
    "completeness_pct": 100
  },
  "error": null
}