Implementation:FlowiseAI Flowise InsertIntoVectorStore

Attribute	Value
Sources	packages/ui/src/api/documentstore.js:L22
Domains	Document_Store_Ingestion
Last Updated	2026-02-12 14:00 GMT

Overview

InsertIntoVectorStore is the API client function that triggers the upsert of processed document chunks into a configured vector store. It sends the complete embedding, vector store, and optional record manager configuration to the server, which handles embedding generation, vector storage, and deduplication.

Code Reference

Source Location

• File: packages/ui/src/api/documentstore.js, Line 22
• Repository: FlowiseAI/Flowise

Signature

const insertIntoVectorStore = (body) => client.post(`/document-store/vectorstore/insert`, body)

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

Import

import documentStoreApi from '@/api/documentstore'

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

I/O Contract

Inputs

Parameter	Type	Required	Description
body	object	Yes	The complete upsert configuration
body.storeId	string	Yes	The document store identifier containing the chunks to upsert
body.docId	string	No	Specific document/loader ID to upsert (omit to upsert all chunks in the store)
body.isStrictSave	boolean	Yes	Whether to persist the vector store configuration alongside the upsert
body.embeddingName	string	Yes	The embedding provider component name (e.g., openAIEmbeddings)
body.embeddingConfig	object	Yes	Embedding provider configuration (model name, credentials, parameters)
body.vectorStoreName	string	Yes	The vector store provider component name (e.g., pinecone, chroma)
body.vectorStoreConfig	object	Yes	Vector store provider configuration (index, namespace, connection details)
body.recordManagerName	string	No	The record manager component name (e.g., postgresRecordManager)
body.recordManagerConfig	object	No	Record manager configuration (table name, cleanup mode, connection details)

Outputs

Field	Type	Description
data	object	The upsert result summary
data.numAdded	number	Number of new vectors added to the store
data.numUpdated	number	Number of existing vectors updated
data.numSkipped	number	Number of unchanged vectors skipped (when using record manager)
data.numDeleted	number	Number of vectors deleted (when record manager cleanup removes stale entries)

The function returns a Promise<{data: {numAdded: number, numUpdated: number, numSkipped: number, numDeleted: number}}>.

Usage Examples

Basic Upsert with OpenAI and Pinecone

import documentStoreApi from '@/api/documentstore'

const upsertToPinecone = async (storeId) => {
    const body = {
        storeId,
        isStrictSave: true,
        embeddingName: 'openAIEmbeddings',
        embeddingConfig: {
            modelName: 'text-embedding-ada-002',
            credential: 'openai-api-key-id'
        },
        vectorStoreName: 'pinecone',
        vectorStoreConfig: {
            pineconeIndex: 'my-index',
            pineconeNamespace: 'product-docs'
        }
    }

    const response = await documentStoreApi.insertIntoVectorStore(body)
    const { numAdded, numUpdated, numSkipped, numDeleted } = response.data

    console.log(`Upsert complete:`)
    console.log(`  Added: ${numAdded}`)
    console.log(`  Updated: ${numUpdated}`)
    console.log(`  Skipped: ${numSkipped}`)
    console.log(`  Deleted: ${numDeleted}`)

    return response.data
}

Upsert with Record Manager for Incremental Updates

import documentStoreApi from '@/api/documentstore'

const incrementalUpsert = async (storeId) => {
    const body = {
        storeId,
        isStrictSave: true,
        embeddingName: 'openAIEmbeddings',
        embeddingConfig: {
            modelName: 'text-embedding-ada-002',
            credential: 'openai-key'
        },
        vectorStoreName: 'chroma',
        vectorStoreConfig: {
            collectionName: 'my-collection',
            chromaURL: 'http://localhost:8000'
        },
        recordManagerName: 'postgresRecordManager',
        recordManagerConfig: {
            tableName: 'doc_records',
            cleanup: 'incremental',
            credential: 'postgres-conn'
        }
    }

    const response = await documentStoreApi.insertIntoVectorStore(body)
    console.log(`Incremental upsert: ${response.data.numAdded} new, ${response.data.numSkipped} unchanged`)
    return response.data
}

Upsert a Single Document

import documentStoreApi from '@/api/documentstore'

const upsertSingleDoc = async (storeId, docId, embeddingConfig, vectorStoreConfig) => {
    const body = {
        storeId,
        docId,  // Only upsert chunks from this specific document
        isStrictSave: false,
        embeddingName: 'openAIEmbeddings',
        embeddingConfig,
        vectorStoreName: 'pinecone',
        vectorStoreConfig
    }

    const response = await documentStoreApi.insertIntoVectorStore(body)
    console.log(`Document ${docId}: ${response.data.numAdded} added, ${response.data.numUpdated} updated`)
    return response.data
}

Related Pages

• Principle:FlowiseAI_Flowise_Vector_Store_Upsert
• Heuristic:FlowiseAI_Flowise_Document_Loader_Bypass_Optimization
• Implementation:FlowiseAI_Flowise_GetVectorStoreProviders -- Fetching available providers for configuration
• Implementation:FlowiseAI_Flowise_GetFileChunks -- Reviewing chunks before upsert
• Implementation:FlowiseAI_Flowise_QueryVectorStore -- Testing retrieval after upsert