Upload a document

Endpoint to upload documents. Documents are saved to the persistent storage, parsed, chunked and embedded. The file is indexed as a document in the SQL database, with its chunks, and related to its owner, company and collection for permission application. To specify a workspace, set ‘collection_type’ to one of the following:

• ‘private’ for a personal workspace
• ‘company’ for a company workspace
• ‘workspace’ for other workspaces If ‘workspace’ is specified, ‘workspace_id’ must be provided. If ‘collection_type’ is not specified, it defaults to ‘company’.

POST /api/v2/files

Try it

Upload a document

cURL

curl --request POST \
  --url https://paradigm.lighton.ai/api/v2/files \
  --header 'Authorization: Bearer <token>' \
  --header 'Content-Type: application/json' \
  --data '
{
  "file": "@document.pdf",
  "collection_type": "workspace",
  "workspace_id": 1
}
'

201

{
  "id": 1,
  "object": "list",
  "bytes": 20032,
  "created_at": 1631533200,
  "filename": "document.pdf",
  "purpose": "documents",
  "status": "embedded"
}

Authorizations

Authorization · string · header · required

Bearer token authentication

Body

Request serializer for file upload.

file · string<uri> · required

title · string | null

Maximum string length: 255

filename · string | null

Maximum string length: 512

collection · integer | null

owner · integer

Owner of the file (optional)

company · integer

Company to associate the file with (optional)

content · string | null

file_type · unknown

collection_type · any

Type of collection to store the file in (private, company, or workspace)

• private - private
• company - company
• workspace - workspace

workspace_id · integer

Workspace ID when collection_type is 'workspace'

chunk_size · integer

Size of chunks for document processing (optional)

chunk_overlap · number<double>

Overlap between chunks (optional)

OCR_AGENT · string

OCR agent to use for processing (optional)

OCR_COMPLETE_DOC_TABLE_EXTRACTION · boolean | null

Enable complete document table extraction (optional)

OCR_HI_RES_MODEL_NAME · string

High resolution OCR model name (optional)

OCR_STRATEGY · string

OCR processing strategy (optional)

OCR_TIMEOUT · integer

OCR processing timeout in seconds (optional)

OCR_URL · string

OCR service URL (optional)

IS_VISUAL_INGESTION_ON · boolean | null

Enable visual ingestion processing (optional)

DEFAULT_PIPELINE · string

Default processing pipeline to use (optional)

Response

201

application/json

Document uploaded successfully

Response serializer for file upload results.

id · integer · required

Unique identifier for the uploaded file

object · string · required

Object type, always 'file'

bytes · integer · required

Size of the file in bytes

created_at · integer · required

Unix timestamp of when the file was created

filename · string · required

Original filename of the uploaded file

purpose · string · required

Purpose of the file, always 'documents'

status · string · required

Processing status of the file