Skip to main content

​
Response Formats

The anyformat API supports multiple response formats to accommodate different use cases and integration patterns.

​
Paginated List Responses

Endpoints that return multiple items use paginated responses: 
{
  "count": 25,
  "total_pages": 3,
  "next": "https://api.anyformat.ai/workflows/?page=2",
  "previous": null,
  "results": [
    {
      "id": "550e8400-e29b-41d4-a716-446655440000",
      "name": "Invoice Data Extraction",
      "description": "Extracts invoice data from PDF documents",
      "created_at": "2024-01-15T10:30:00Z",
      "updated_at": "2024-01-20T14:22:00Z"
    }
  ]
}
FieldDescription
countTotal number of items across all pages
total_pagesNumber of pages available
nextURL for the next page (null if last page)
previousURL for the previous page (null if first page)
resultsArray of items for the current page

​
Extraction Results Output Formats

The workflow results endpoint (/workflows/{id}/results/) supports three output formats:
FormatParameterDescription
CSVoutput_format=csvTabular data (default), ideal for spreadsheets
JSONoutput_format=jsonStructured data with full metadata
JSON Linesoutput_format=jsonlOne JSON object per line, ideal for streaming
# CSV (default)
curl "https://api.anyformat.ai/workflows/{id}/results/" \
  -H "x-api-key: YOUR_API_KEY"

# JSON
curl "https://api.anyformat.ai/workflows/{id}/results/?output_format=json" \
  -H "x-api-key: YOUR_API_KEY"

# JSON Lines
curl "https://api.anyformat.ai/workflows/{id}/results/?output_format=jsonl" \
  -H "x-api-key: YOUR_API_KEY"

​
Consistent Formatting with as_lists=true

The as_lists=true parameter transforms extracted fields to a consistent array format, making automated processing easier.
The as_lists=true parameter only works with output_format=json. It is ignored for CSV and JSONL formats.

​
Standard Response

{
  "id": "362963",
  "status": "processed",
  "processed_at": "2024-01-15T10:30:00Z",
  "results": {
    "invoice_id": {
      "value": "13152556",
      "evidence": [{"text": "Invoice #13152556", "page_number": 1}],
      "confidence": 95
    },
    "vendor_name": {
      "value": "Acme Corporation",
      "evidence": [{"text": "From: Acme Corporation", "page_number": 1}],
      "confidence": 92
    },
    "line_items": [
      {
        "description": {"value": "Consulting", "confidence": 90},
        "amount": {"value": 500.00, "confidence": 95}
      },
      {
        "description": {"value": "Support", "confidence": 88},
        "amount": {"value": 250.00, "confidence": 92}
      }
    ]
  }
}

​
With as_lists=true

{
  "id": "362963",
  "status": "processed",
  "processed_at": "2024-01-15T10:30:00Z",
  "results": {
    "invoice_id": [
      {
        "value": "13152556",
        "evidence": [{"text": "Invoice #13152556", "page_number": 1}],
        "confidence": 95
      }
    ],
    "vendor_name": [
      {
        "value": "Acme Corporation",
        "evidence": [{"text": "From: Acme Corporation", "page_number": 1}],
        "confidence": 92
      }
    ],
    "line_items": [
      {
        "description": {"value": "Consulting", "confidence": 90},
        "amount": {"value": 500.00, "confidence": 95}
      },
      {
        "description": {"value": "Support", "confidence": 88},
        "amount": {"value": 250.00, "confidence": 92}
      }
    ]
  }
}
Key differences:
  • Metadata fields (id, status, processed_at) remain unchanged
  • Single-value fields (invoice_id, vendor_name) are wrapped in arrays
  • Already-array fields (line_items) remain unchanged

​
When to Use Each Format

Use CaseRecommended Format
Spreadsheet importoutput_format=csv
Standard API integrationoutput_format=json
Automated data pipelinesoutput_format=json&as_lists=true
Streaming/large datasetsoutput_format=jsonl

​
Supported Endpoints

​
Endpoints supporting as_lists=true

EndpointDescription
GET /workflows/{id}/results/?output_format=json&as_lists=trueWorkflow extraction results
GET /jobs/{id}/?as_lists=trueIndividual job results

​
Endpoints where as_lists=true has no effect

EndpointReason
GET /workflows/Returns workflow metadata, not extraction data
GET /workflows/{id}/Returns single workflow details
Any endpoint returning only metadataNo extracted data fields to transform

​
Result Field Structure

Each extracted field in the results contains: 
{
  "field_name": {
    "value": "extracted_value",
    "confidence": 95,
    "evidence": [
      {
        "text": "Snippet from document",
        "page_number": 1
      }
    ],
    "verification_status": "not_verified",
    "value_unit": null
  }
}
PropertyDescription
valueThe extracted value
confidenceConfidence score (0-100)
evidenceArray of source locations in the document
verification_statusWhether the value has been manually verified
value_unitUnit of measurement (e.g., “USD” for currency)