Embedding vectors

Produce numeric representations for an application-managed index.

No current prepared translation is available. The English source is shown without machine translation.

Call POST /v1/embeddings with an embedding model available to the credential. Keep the model choice and output dimension consistent when populating a search index. This endpoint returns vectors, not a conversational answer.

Request

curl https://api.b2b-llm.com/v1/embeddings \
  -H "Authorization: Bearer $B2B_API_KEY" -H 'Content-Type: application/json' \
  -d '{"model":"embedding-model","input":["Purchase order received","Invoice awaiting review"]}'

embedding-model is an illustrative placeholder, not a guaranteed catalog entry.

Response

{"object":"list","data":[{"object":"embedding","index":0,"embedding":[0.08,-0.12,0.03]},{"object":"embedding","index":1,"embedding":[0.06,0.02,-0.09]}]}

The shortened vectors illustrate structure only. Store the full returned vectors in your application and use index to associate each result with its input. Apply the model's input-size limits rather than truncating text silently. For multiple offline inputs, the Files/Batch interface can execute supported embedding items.