Skip to main content
POST
/
api
/
v1
/
document-store
/
upsert
/
{storeId}
Upsert to Document Store
curl --request POST \
  --url https://api.example.com/api/v1/document-store/upsert/{storeId} \
  --header 'Content-Type: application/json' \
  --data '
{
  "files": [
    {}
  ],
  "loaderId": "<string>",
  "chunkSize": 123,
  "chunkOverlap": 123,
  "metadata": {}
}
'
{
  "400": {},
  "401": {},
  "403": {},
  "404": {},
  "412": {},
  "500": {},
  "numAdded": 123,
  "numUpdated": 123,
  "numSkipped": 123,
  "numDeleted": 123,
  "addedDocs": [
    {}
  ]
}
Upload and process documents into a document store. This endpoint handles file uploads, chunking, embedding, and upserting to the vector store.

​
Path Parameters

​
storeId
string
required
The unique identifier of the document store to upsert documents into

​
Request Body

This endpoint accepts multipart/form-data for file uploads.
​
files
array
required
Array of files to upload. Supports PDF, TXT, DOCX, CSV, and other formats depending on configured loaders
​
loaderId
string
Specific loader ID to use for processing files
​
chunkSize
number
Size of text chunks (default varies by loader)
​
chunkOverlap
number
Overlap between chunks (default varies by loader)
​
metadata
object
Additional metadata to attach to all documents

​
Response

​
numAdded
number
Number of new documents added
​
numUpdated
number
Number of existing documents updated
​
numSkipped
number
Number of documents skipped
​
numDeleted
number
Number of documents deleted
​
addedDocs
array
IDs of added documents

​
Example Request

curl -X POST \
  'https://your-flowise-instance.com/api/v1/document-store/upsert/store-123' \
  -H 'Authorization: Bearer YOUR_API_KEY' \
  -F 'files=@/path/to/document.pdf' \
  -F 'files=@/path/to/guide.txt' \
  -F 'chunkSize=1000' \
  -F 'chunkOverlap=200'

​
Example Request - With Metadata

curl -X POST \
  'https://your-flowise-instance.com/api/v1/document-store/upsert/store-123' \
  -H 'Authorization: Bearer YOUR_API_KEY' \
  -F 'files=@/path/to/document.pdf' \
  -F 'metadata={"category":"technical","version":"1.0"}'

​
Example Response

{
  "numAdded": 45,
  "numUpdated": 0,
  "numSkipped": 0,
  "numDeleted": 0,
  "addedDocs": [
    "doc-001",
    "doc-002",
    "doc-003"
  ]
}

​
Supported File Types

Depending on configured loaders, supported formats include:
  • PDF (.pdf)
  • Text (.txt)
  • Word (.docx)
  • Markdown (.md)
  • CSV (.csv)
  • JSON (.json)
  • HTML (.html)

​
Processing Steps

  1. Upload - Files are uploaded and validated
  2. Load - Appropriate loader processes the files
  3. Chunk - Content is split into chunks
  4. Embed - Chunks are embedded using configured embedding model
  5. Upsert - Embeddings are upserted to vector store
  6. Index - Documents are indexed for retrieval

​
Chunking Strategies

  • Fixed Size - Split by character count
  • Recursive - Split by paragraphs, sentences, then characters
  • Semantic - Split by semantic meaning

​
Error Responses

​
400
error
Bad Request - Invalid file format or missing required fields
​
401
error
Unauthorized - Invalid or missing API key
​
403
error
Forbidden - Insufficient permissions to upsert documents
​
404
error
Not Found - Document store not found
​
412
error
Precondition Failed - Missing store ID or workspace ID
​
500
error
Internal Server Error - Error processing documents

Build docs developers (and LLMs) love

Get started for freeTalk to us