Upload a Document to a Corpus

Add a document to a collection from a file, text, or URL. The document is queued for
indexing and becomes searchable in the background. A 200 response means the document
was accepted, not yet indexed.

Note:

On corpora with PII redaction turned on, personal information is removed from uploaded
files and text before they are stored, and the stored copy is a .txt file.

Current Version:

2.0.0

application/json

The upload was handled. Check success: a rejected or failed upload also returns 200, with success false, step failed, and id and access_url null when no document was created.

Bodyapplication/json

curl --location --globoff 'https:///indexing-service/api/v2/corpora//documents' \ --header 'Ocp-Apim-Subscription-Key: ' \ --header 'Authorization: Bearer ' \ --form 'user_id="42"' \ --form 'content_type="file"' \ --form 'source=@"report.pdf"' \ --form 'upload_from="web"' \ --form 'access_url="https://example.com/document.pdf"' \ --form 'data_id="5521"' \ --form 'tags="finance"' \ --form 'tags="q2"' \ --form 'step="pending"' \ --form 'refresh="false"' \ --form 'etag="d41d8cd98f00b204e9800998ecf8427e"' \ --form 'sitemap_id="123"'

{ "success": true, "message": "Document uploaded and extraction task queued", "data": { "id": 5521, "source": "document1715170000.pdf", "access_url": "https://storage.example.com/citations/document1715170000.pdf", "step": "pending" } }

Upload a Document to a Corpus

Request

Responses