Document Extracted Event

Webhook event fired when document processing completes, whether successful or failed.

The DOCUMENT_EXTRACTED event fires when document processing completes, whether the extraction was successful or failed. This provides immediate feedback on extraction job status.

When It Triggers

Successful extraction -- When a document is successfully processed and data is extracted
Failed extraction -- When document processing fails due to errors or timeouts

Typically fires within 30-120 seconds after calling the process-file or process-file-async endpoint, depending on document complexity.

On a model with at least one validation rule, the rules are evaluated before this event is published, so a read triggered by it sees the document's final rule results -- validations.rules and the per-cell RULE entries are already the verdict for this extraction. A model with no rules has nothing to evaluate, and the event is published as soon as the extraction completes.

When rule evaluation fails, this event is published first and DOCUMENT_VALIDATION_FAILED follows it; the document's validation status reads error from the moment this event goes out, so a read triggered by it is still conclusive.

Custom validations are not covered either way: scripts and validator agents are dispatched after the event and settle later, announced by DOCUMENT_VALIDATED.

Payload Structure

Successful Extraction

Response200

{
  "eventId": "evt_success_456",
  "eventType": "DOCUMENT_EXTRACTED",
  "timestamp": "2024-01-15T10:30:45Z",
  "organizationId": "org_xyz789",
  "projectId": "proj_abc123",
  "deploymentId": "dep_123abc",
  "viaDeploymentId": "dpl_prod01",
  "deploymentName": "production",
  "schemaVersionId": "sv_7",
  "extractionSchemaId": "schema_456def",
  "fileId": "file_789ghi",
  "jobId": "job_012jkl",
  "data": {
    "status": "success",
    "errorMessage": null
  }
}

Failed Extraction

Response200

{
  "eventId": "evt_failed_789",
  "eventType": "DOCUMENT_EXTRACTED",
  "timestamp": "2024-01-15T10:32:15Z",
  "organizationId": "org_xyz789",
  "projectId": "proj_abc123",
  "deploymentId": "dep_123abc",
  "viaDeploymentId": "dpl_prod01",
  "deploymentName": "production",
  "schemaVersionId": "sv_7",
  "extractionSchemaId": "schema_456def",
  "fileId": "file_bad123",
  "jobId": "job_error456",
  "data": {
    "status": "failed",
    "errorMessage": "Document format not supported"
  }
}

Payload Fields

Field	Type	Description
`eventId`	string	Unique event identifier. Use for idempotency.
`eventType`	string	Always `"DOCUMENT_EXTRACTED"`
`timestamp`	ISO 8601	When the extraction completed (UTC)
`organizationId`	string	Your organization identifier
`projectId`	string	The project the document belongs to
`deploymentId`	string	The ID of the model that processed this document
`viaDeploymentId`	string \| null	The ID of the deployment the request arrived through. `null` when the request named the model directly
`deploymentName`	string \| null	The name of that deployment, such as `"production"`. `null` whenever `viaDeploymentId` is `null`
`schemaVersionId`	string \| null	The ID of the frozen schema version the run resolved to. `null` when the deployment tracks the draft schema
`extractionSchemaId`	string	The extraction schema ID of that model
`fileId`	string	The processed file identifier
`jobId`	string	The extraction job identifier
`data.status`	`"success"` \| `"failed"`	Extraction result
`data.errorMessage`	string \| null	Error description if failed, otherwise null

When several deployments point at one model, viaDeploymentId tells you which of them served the traffic.

Handling the Event

function handleDocumentExtracted(event) {
const { eventId, fileId, jobId, data } = event

if (data.status === "success") {
  console.log(`Document ${fileId} extracted successfully (job: ${jobId})`)

  fetchFileData(fileId).then(fileData => {
    processExtractedData(fileData)
  })
} else {
  console.error(`Document ${fileId} extraction failed: ${data.errorMessage}`)
}
}

Retrieving Extracted Data

The webhook payload contains metadata only. To get the actual extracted data, call the get-file-data endpoint using the fileId from the payload.