Skip to main content
POST
/
generate-document
Generate document field values from evidence
curl --request POST \
  --url https://ai.platform.arb.inc/generate-document \
  --header 'Authorization: Bearer <token>' \
  --header 'Content-Type: application/json' \
  --data '
{
  "documents": [
    {
      "gsUrl": "gs://platform-evidence/2025/invoice-2025-001.pdf",
      "contentType": "application/pdf",
      "description": "Invoice showing unpaid parking fees"
    }
  ],
  "documentType": "statementOfClaim",
  "context": "Focus on extracting party information and claim amounts",
  "referenceDocuments": [
    {
      "documentType": "statementOfClaim",
      "gsUrl": "gs://platform-cases/2025/case-123/statement-of-claim.pdf"
    }
  ]
}
'
{
  "documentType": "statementOfClaim",
  "fields": {
    "claimantName": "Vanguard Parking Solutions",
    "claimantAddress": "3389 SHERIDAN ST BOX",
    "claimantCity": "HOLLYWOOD",
    "claimantState": "FL",
    "claimantZipcode": "33021"
  }
}

Authorizations

​
Authorization
string
header
required

access token

Body

application/json
​
documents
object[]
required

Evidence or source documents to analyze

Minimum array length: 1
​
documentType
string
required

Document type identifier matching platform document schema

Required string length: 1 - 100
Example:

"statementOfClaim"

​
context
string

Optional user-provided context to guide document generation

Maximum string length: 10000
Example:

"Focus on extracting party information and claim amounts"

​
referenceDocuments
object[]

Previously filed documents to reference for context

Maximum array length: 5

Response

200 - application/json

Generated document fields

​
documentType
string
Example:

"statementOfClaim"

​
fields
object

Map of field IDs to extracted values

Example:
{
  "claimantName": "Vanguard Parking Solutions",
  "claimantAddress": "3389 SHERIDAN ST BOX",
  "claimantCity": "HOLLYWOOD",
  "claimantState": "FL",
  "claimantZipcode": "33021"
}