Implementation:FlowiseAI Flowise PreviewChunks

Attribute	Value
Sources	packages/ui/src/api/documentstore.js:L17
Domains	Document_Store_Ingestion
Last Updated	2026-02-12 14:00 GMT

Overview

PreviewChunks is the API client function that sends a document loader and splitter configuration to the server for trial processing, returning a limited set of sample chunks without persisting them. This enables users to evaluate their configuration before committing to full document processing.

Code Reference

Source Location

• File: packages/ui/src/api/documentstore.js, Line 17
• Repository: FlowiseAI/Flowise

Signature

const previewChunks = (body) => client.post('/document-store/loader/preview', body)

This function sends a POST request to the /api/v1/document-store/loader/preview endpoint via the configured axios client instance.

Import

import documentStoreApi from '@/api/documentstore'

The function is accessed as documentStoreApi.previewChunks(body) after importing the default export from the documentstore API module.

I/O Contract

Inputs

Parameter	Type	Required	Description
body	object	Yes	The complete loader and splitter configuration
body.storeId	string	Yes	The document store identifier
body.loaderId	string	Yes	The loader instance identifier (empty string for new loaders)
body.loaderName	string	Yes	The loader component name (e.g., pdfLoader)
body.loaderConfig	object	Yes	Loader-specific configuration (file references, URLs, API keys, etc.)
body.splitterId	string	No	The text splitter component name (e.g., recursiveCharacterTextSplitter)
body.splitterConfig	object	No	Splitter-specific configuration (chunkSize, chunkOverlap, separators)
body.credential	string	No	Credential identifier for loaders requiring authentication
body.previewChunkCount	number	Yes	Maximum number of preview chunks to return (typically 20)

Outputs

Field	Type	Description
data	object	The preview result object
data.totalChunks	number	Total number of chunks that would be produced from full processing
data.chunks	Chunk[]	Array of sample chunk objects (limited by previewChunkCount)
data.chunks[].pageContent	string	The text content of the chunk
data.chunks[].metadata	object	Metadata associated with the chunk (source, page, line numbers, etc.)
data.previewChunkCount	number	The number of chunks actually returned in this preview

The function returns a Promise<{data: {totalChunks: number, chunks: Chunk[], previewChunkCount: number}}> from the axios response.

Usage Examples

Basic Preview with PDF Loader

import documentStoreApi from '@/api/documentstore'

const previewPdfChunks = async (storeId, fileId) => {
    const body = {
        storeId,
        loaderId: '',
        loaderName: 'pdfLoader',
        loaderConfig: { pdfFile: fileId },
        splitterId: 'recursiveCharacterTextSplitter',
        splitterConfig: {
            chunkSize: 1000,
            chunkOverlap: 200
        },
        previewChunkCount: 20
    }

    const response = await documentStoreApi.previewChunks(body)
    const { totalChunks, chunks } = response.data

    console.log(`Preview: ${chunks.length} of ${totalChunks} total chunks`)
    chunks.forEach((chunk, i) => {
        console.log(`--- Chunk ${i + 1} ---`)
        console.log(`Content: ${chunk.pageContent.substring(0, 200)}...`)
        console.log(`Metadata:`, chunk.metadata)
    })

    return response.data
}

Preview Without a Splitter

import documentStoreApi from '@/api/documentstore'

// Preview using only the loader (no text splitting)
const previewRawDocuments = async (storeId) => {
    const body = {
        storeId,
        loaderId: '',
        loaderName: 'textFile',
        loaderConfig: { textFile: 'uploaded-file-ref' },
        previewChunkCount: 10
    }

    const response = await documentStoreApi.previewChunks(body)
    return response.data
}

Iterative Configuration Tuning

import documentStoreApi from '@/api/documentstore'

const tuneChunkSize = async (storeId, loaderConfig) => {
    const chunkSizes = [500, 1000, 1500, 2000]

    for (const chunkSize of chunkSizes) {
        const body = {
            storeId,
            loaderId: '',
            loaderName: 'pdfLoader',
            loaderConfig,
            splitterId: 'recursiveCharacterTextSplitter',
            splitterConfig: { chunkSize, chunkOverlap: Math.floor(chunkSize * 0.2) },
            previewChunkCount: 5
        }

        const response = await documentStoreApi.previewChunks(body)
        console.log(`Chunk size ${chunkSize}: ${response.data.totalChunks} total chunks`)
    }
}

Related Pages

• Principle:FlowiseAI_Flowise_Chunk_Preview
• Implementation:FlowiseAI_Flowise_GetDocumentLoaders -- Fetching available loader definitions
• Implementation:FlowiseAI_Flowise_GetNodesByCategory -- Fetching available splitter definitions
• Implementation:FlowiseAI_Flowise_SaveProcessingLoader -- Committing to full processing after preview