APIOfficial documentation7 min readAPI

FastGPT Dataset API Data Structure Reference

This reference document outlines the official data structure for dataset entries used in the FastGPT OpenAPI.

Overview

This reference document outlines the official data structure for dataset entries used in the FastGPT OpenAPI. These structures are required for programmatically creating, updating, and managing dataset data, ensuring consistent formatting for vector embedding and retrieval operations.

Core Data Structure

The following table lists all valid fields for a FastGPT dataset data entry, including their data type, functional description, and requirement status for API requests:

FieldTypeDescriptionRequired
teamIdStringTeam ID
tmbIdStringMember ID
datasetIdStringDataset ID
collectionIdStringCollectionID
qStringPrimary data
aStringAuxiliary data
fullTextTokenStringTokenization
indexesIndex[]Vector indexes
updateTimeDateUpdate time
chunkIndexNumberChunk index

Index Sub-Structure

Each dataset data entry can include up to 5 custom vector indexes. The following table defines the valid fields for an individual index entry:

FieldTypeDescriptionRequired
typeStringOptional index types: default-default index; custom-custom index; summary-summary index; question-question index; image-image index
dataIdStringAssociated vector ID. Pass this ID when updating data for incremental updates instead of full updates
textStringText content

Per the official specification, if the type field is not provided for an index, it defaults to custom. A default index will also be created based on the q and a fields of the parent data entry unless a default index is explicitly included in the indexes array, in which case no additional default index is generated.

Source: FastGPT official source

Applicability and version scope

Use this page for the documented API scenario. Confirm the FastGPT, dependency, API, and deployment versions in the official source before applying a change.

Safety guardrails

Use [REDACTED_CREDENTIAL] for credentials and private data. Confirm the documented environment and version before review.

Rollback guidance

Restore the prior technical-content authority snapshot. Restore saved configuration and data snapshots, then repeat the smallest verification scenario.