Implementation:Langgenius Dify CreateEmptyDataset

Knowledge Sources	Domains	Last Updated
Dify	RAG, Knowledge_Management, Frontend	2026-02-12 00:00 GMT

Overview

Description

createEmptyDataset is a frontend service function that creates a new, empty knowledge base (dataset) in Dify. It issues a POST request to the /datasets endpoint with only a name, and the backend provisions the full dataset resource with default configuration for indexing, permissions, embedding model, and retrieval model.

This function represents the simplest entry point into the Knowledge Base Management workflow. Once the dataset is created, documents can be uploaded, chunked, indexed, and queried within it.

Usage

• Call createEmptyDataset when the user initiates the "Create Knowledge Base" action from the UI.
• The returned DataSet object provides the id needed to add documents, configure settings, or attach the dataset to applications.
• Typically followed by a call to createFirstDocument or createDocument to populate the newly created dataset.

Code Reference

Source Location

web/service/datasets.ts, lines 84--86.

Signature

export const createEmptyDataset = ({ name }: { name: string }): Promise<DataSet> => {
  return post<DataSet>('/datasets', { body: { name } })
}

Import

import { createEmptyDataset } from '@/service/datasets'

I/O Contract

Inputs

Parameter	Type	Required	Description
name	string	Yes	Human-readable name for the new dataset.

Outputs

Returns Promise<DataSet>. Key fields of the DataSet type:

Field	Type	Description
id	string	Unique identifier for the created dataset.
name	string	Name of the dataset as provided in the request.
indexing_status	DocumentIndexingStatus	Current indexing status of the dataset.
permission	DatasetPermission	Access control level: only_me, all_team_members, or partial_members.
doc_form	ChunkingMode	Default chunking mode: text_model, qa_model, or hierarchical_model.
runtime_mode	'rag_pipeline' ¦ 'general'	Whether the dataset operates as a standard knowledge base or a RAG pipeline.
embedding_model	string	Name of the embedding model assigned to the dataset.
embedding_model_provider	string	Provider of the embedding model.
retrieval_model	RetrievalConfig	Default retrieval configuration including search method, top_k, and score threshold.

Usage Examples

Creating a new knowledge base

import { createEmptyDataset } from '@/service/datasets'

const dataset = await createEmptyDataset({ name: 'Product Documentation' })
console.log(dataset.id) // Use this ID for subsequent document uploads

Creating a dataset and immediately adding a document

import { createEmptyDataset, createDocument } from '@/service/datasets'

const dataset = await createEmptyDataset({ name: 'FAQ Knowledge Base' })
const docResponse = await createDocument({
  datasetId: dataset.id,
  body: {
    data_source: { type: 'upload_file', info_list: { data_source_type: 'upload_file', file_info_list: { file_ids: [fileId] } } },
    doc_form: 'text_model',
    doc_language: 'English',
    process_rule: processRule,
    retrieval_model: dataset.retrieval_model,
    embedding_model: dataset.embedding_model,
    embedding_model_provider: dataset.embedding_model_provider,
  },
})

Related Pages

• Principle:Langgenius_Dify_Dataset_Creation