Vectara · Schema

Vectara Document

A Vectara document is either a structured (Core) document with explicit parts, or an unstructured file that Vectara parses into parts.

AIAgentsCorporaEmbeddingsEnterprise SearchGenerative AIGrounded GenerationHallucination DetectionLLMMCPRAGRetrievalSearchSemantic SearchVector Search

Properties

Name Type Description
id string Caller-assigned unique identifier within the corpus.
type string Whether the document was added via Core indexing or file upload.
metadata object Document-level metadata used for filtering and presentation.
document_parts array Ordered list of document parts (sections, chunks) indexed for retrieval.
View JSON Schema on GitHub

JSON Schema

vectara-document-schema.json Raw ↑
{
  "$schema": "https://json-schema.org/draft/2020-12/schema",
  "$id": "https://raw.githubusercontent.com/api-evangelist/vectara/main/json-schema/vectara-document-schema.json",
  "title": "Vectara Document",
  "description": "A Vectara document is either a structured (Core) document with explicit parts, or an unstructured file that Vectara parses into parts.",
  "type": "object",
  "required": ["id"],
  "properties": {
    "id": {
      "type": "string",
      "description": "Caller-assigned unique identifier within the corpus."
    },
    "type": {
      "type": "string",
      "enum": ["core", "structured"],
      "description": "Whether the document was added via Core indexing or file upload."
    },
    "metadata": {
      "type": "object",
      "additionalProperties": true,
      "description": "Document-level metadata used for filtering and presentation."
    },
    "document_parts": {
      "type": "array",
      "description": "Ordered list of document parts (sections, chunks) indexed for retrieval.",
      "items": {
        "type": "object",
        "required": ["text"],
        "properties": {
          "text": { "type": "string" },
          "metadata": {
            "type": "object",
            "additionalProperties": true
          },
          "context": { "type": "string" }
        }
      }
    }
  }
}