A document with a unique identifier and field values. Fields named in the index schema are validated against it; any other field is stored as filterable metadata. Every document must carry at least one schema field and every schema field the index marks required; a document with only _id and metadata is rejected. Limits: 2 MB per document and per request, 100 KB per string schema field value (whether or not its text is stored), and 10,000 tokens per full-text-search field value.
A string field that declares embed or sparse_embed takes its text as a string, and Pinecone derives its vectors; a caller-supplied vector on such a field is rejected. A field with embed is required on every upsert.
A document with a unique identifier and field values. Fields named in the index schema are validated against it; any other field is stored as filterable metadata. Every document must carry at least one schema field and every schema field the index marks required; a document with only
_idand metadata is rejected. Limits: 2 MB per document and per request, 100 KB perstringschema field value (whether or not its text is stored), and 10,000 tokens per full-text-search field value.A
stringfield that declaresembedorsparse_embedtakes its text as a string, and Pinecone derives its vectors; a caller-supplied vector on such a field is rejected. A field withembedis required on every upsert.Export
DocumentRecord