The total number of tokens Pinecone embedded for the request, summed over every embedding call it made, dense and sparse, across all fields and models. Documents aren't deduplicated by _id within an upsert, so every entry's text is embedded and counted, including entries that repeat an _id.
Usage information for the
upsert_documentsandupdate_documentsoperations.Present only when the request embedded text, that is, when it wrote text to a field whose schema declares
embedorsparse_embed. A request that embedded nothing returns nousage.Export
DocumentWriteUsage