Indexing

These endpoints add documents to a docset and let you automate ingestion pipelines. A docset is indexed from a source — S3 (default), SharePoint, or Google Drive.

Index a docset

POST /index

Field	Type	Required	Description
`docSetName`	string	✅	The docset to index into.
`sourceType`	enum		`S3` (default), `SHAREPOINT`, or `GOOGLEDRIVE`.
`credentials`	object		Source credentials; shape depends on `sourceType`. Omit for S3 when the API is configured with bucket access.
`indexOptions`	object		`{ targetIndexTypes?, forceReindex? }`.
`async`	boolean		Queue the job (`true`, default) or block until done (`false`).
`accountName`	string		Optional account/tenant label.
`summaryModel`	string		LLM used for document summarization.


curl -X POST "$API_URL/index" \
  -H "Authorization: Bearer $API_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{ "docSetName": "company-policies", "sourceType": "S3" }'

By default indexing runs asynchronously and returns a job id; poll /index/progress to follow it:


{ "message": "Indexing queued", "docSetName": "company-policies", "jobId": "..." }

For non-S3 sources, the API probes the source synchronously before accepting the request and returns 422 if it can’t connect or finds 0 files while documents already exist. Check the source first to surface credential problems early.

Check a source before indexing

POST /index/check-source

Validates connectivity and reports how many supported files the source has, without indexing anything.


curl -X POST "$API_URL/index/check-source" \
  -H "Authorization: Bearer $API_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{ "docSetName": "company-policies", "sourceType": "S3" }'


{ "sourceFileCount": 12, "existingDocumentCount": 0, "message": "Found 12 supported files." }

Track progress

GET /index/progress/{docSetName}

Returns document status counts and an isIndexing flag. For SharePoint, you can scope progress to a folder by passing siteId, driveId, and folderPath together.


curl "$API_URL/index/progress/company-policies" \
  -H "Authorization: Bearer $API_TOKEN"


{
  "pendingCount": 3,
  "indexingCount": 1,
  "completedCount": 8,
  "errorCount": 0,
  "isIndexing": true
}

Other indexing operations

Operation	Endpoint
Cancel an in-progress run	`POST /index/cancel/{docSetName}`
Retry a single failed document	`POST /index/retry/{documentId}`
Delete docsets	`DELETE /index/docsets` (body: `{ "docSetNames": [...] }`)

See the API Reference for full request and response shapes.