GET /extracts — Retrieve Verification Results

Call this endpoint after GET /status reports master_status: SUCCESS (or FAILED) to retrieve the full extraction payload. The response contains per-document OCR fields and pairwise cross-check comparisons across all uploaded documents.

Endpoint

GET /v3/verification/extracts

Request headers

Token

string

required

JWT access token obtained from POST /authtoken.

Query parameters

flow_uuid

string

required

The UUID of the completed flow, as returned by POST /flow/init.

Response

`ocr_data` array

Each item in ocr_data corresponds to one uploaded document.

ocr_data

array

List of OCR extraction results, one per uploaded document.

Show ocr_data item

data

object

Document-type-specific extracted fields. The shape of this object varies by document_type (for example, INVOICE, IMEI_CLOSEUP_IMAGE, DELIVERY_ORDER). Each field typically contains a value and a confidence_score.

document_id

string

The document_id you supplied during upload. Use this to match the result back to your internal records.

error_code

string

Machine-readable error code if extraction failed. Common values:

INVALID_DOC — the document was recognised but is invalid (for example, it contains multiple people in an open-box image).
UNRECOGNISED — Atlas could not identify the document type. Empty string when extraction succeeded.

error_reason

string

Human-readable explanation corresponding to error_code (for example, "multiple people", "document not in list"). Empty string when extraction succeeded.

`cross_checks` array

Each item represents a single pairwise field comparison between two documents.

cross_checks

array

Pairwise validation results across all uploaded documents.

Show cross_checks item

source_doc

string

Document type of the source document (for example, "DELIVERY_ORDER").

source_field_name

string

The field name extracted from the source document (for example, "dealer_address").

source_value

string

The extracted text value from the source document field.

target_doc

string

Document type of the target document being compared against (for example, "INVOICE").

target_field_name

string

The field name in the target document (for example, "dealer_address").

target_value

string

The extracted text value from the target document field.

similarity_score

integer

A score from 0 to 100 indicating how closely the two field values match. Higher values indicate stronger agreement between documents.

name

string

A descriptive name for this cross-check (for example, "DEALER_ADDRESS", "CUSTOMER_NAME", "PHOTO_MATCH").

Example

curl --request GET \
  --url "https://api.helloatlas.in/v3/verification/extracts?flow_uuid=c6a2ffd7-eb44-4d16-8b45-32adc2182a8b" \
  --header "Token: eyJhbGciOiJIUzI1NiIsInR5cCI6IkpXVCJ9..."

Error responses

Status	Meaning
`400`	`Missing required ['flow_uuid'] parameters` — the query parameter was not included.
`401`	`Invalid token` — the `Token` header is missing, expired, or invalid.
`409`	`Flow not found` — no flow exists for the provided `flow_uuid`.

Only call this endpoint after confirming master_status is SUCCESS or FAILED via GET /status. Calling /extracts on an in-progress flow may return incomplete results.

Authentication

Document Verification

Chat Based Onboarding

Bulk Service

Address Match

GET /extracts — Retrieve Verification Results

Endpoint

Request headers

Query parameters

Response

`ocr_data` array

`cross_checks` array

Example

Error responses

Authentication

Document Verification

Chat Based Onboarding

Bulk Service

Address Match

​Endpoint

​Request headers

​Query parameters

​Response

​ocr_data array

​cross_checks array

​Example

​Error responses

Endpoint

Request headers

Query parameters

Response

`ocr_data` array

`cross_checks` array

Example

Error responses