Core integration mechanics

New to the API? Start with Quickstart API. This page is a detailed reference for production-grade integrations.

Base URL

Use the stage/cell-specific base URL for your tenant.

Stage and cell mapping: see Concepts
Token URL mapping: see Getting an access token

Step 1 — Create a submission

POST /v2/submission Creates a new submission record and returns:

a unique submission_id
a presigned upload_url for uploading the file bytes

Request headers

Authorization: Bearer <access_token>
Content-Type: application/json

Request body (optional)

{
  "query_id": "your-internal-document-id",
  "pipeline_configuration": "FRAUD_ONLY",
  "enable_decision": false,
  "enable_submission_characteristics": false
}

Field	Type	Default	Notes
`query_id`	string (0–1024)	`null`	Optional correlation ID. Must not contain PII.
`pipeline_configuration`	enum	`FRAUD_ONLY`	Selects which pipeline runs for this submission.
`enable_decision`	boolean	`false`	Enables Adaptive Decision (if available for your tenant).
`enable_submission_characteristics`	boolean	`false`	If `true`, characteristics must be submitted before analysis proceeds.

Pipeline configuration values

Value	Description
`FRAUD_ONLY`	Run fraud analysis (default)
`CLASSIFICATION_ONLY`	Run document classification only
`QUALITY_ONLY`	Run quality analysis only

This enum selects a single pipeline. If you need multiple outputs, retrieve them from the corresponding endpoint(s) or use an agreed configuration provided during provisioning.

Response — 200 OK

{
  "submission_id": "sub_abc123xyz",
  "upload_url": "https://bucket.s3.amazonaws.com/...",
  "submission_characteristics_upload_url": "/v2/submission/sub_abc123xyz/characteristics"
}

Field	Description
`submission_id`	Unique identifier for this submission. Store it — used in all subsequent calls.
`upload_url`	Presigned URL for uploading the document bytes (Step 2).
`submission_characteristics_upload_url`	Present only when submission characteristics are enabled.

Step 2 — Upload the document bytes

Upload the original file bytes to the presigned upload_url using HTTP PUT.

Upload request

UPLOAD_URL="<upload_url_from_step_1>"
FILE_PATH="/path/to/your/document.pdf"

curl --request PUT \
  --url "$UPLOAD_URL" \
  --header "content-type: application/octet-stream" \
  --data-binary @"$FILE_PATH"

Upload rules

Method must be PUT
Content-Type must be application/octet-stream
Upload the original bytes (do not base64 encode or re-render)
The presigned URL expires — upload promptly after creating the submission

If upload fails, validate your outbound allowlist and proxy rules: see Domains to allowlist.

Step 3 — Submit submission characteristics (optional)

PUT /v2/submission/{submission_id}/characteristics If you created the submission with enable_submission_characteristics: true, you must submit characteristics before analysis proceeds.

Do not enable submission characteristics unless you intend to send them for every submission. When enabled, they are required for that submission.

This page does not document the full schema (it is extensive and evolves). See:

Submission characteristics

Response — 204 No Content

Characteristics accepted — analysis will proceed.

Step 4 — Fetch analysis results

Results are returned asynchronously. Fetch them by calling the relevant endpoint(s) for your pipeline and feature configuration.

Polling semantics (important)

A 404 can mean “result not ready yet” or “not found” (e.g., wrong stage/cell, wrong submission_id). If 404 persists, verify you’re using the correct base URL for the same stage/cell where the submission was created.

See Polling for results for recommended exponential backoff and timeouts.

Fraud analysis

GET /v2/submission/{submission_id}/fraud Returns fraud indicators and an overall risk score.

Query parameters

Parameter	Type	Default	Description
`with_metadata`	boolean	`false`	Includes extended indicator metadata (e.g., for UI visualization / localized detections).

Response (example)

{
  "status": "SUCCESS",
  "score": "NORMAL",
  "indicators": [
    {
      "indicator_id": "in_transaction_cluster",
      "type": "RISK",
      "category": "...",
      "title": "...",
      "origin": "fraud"
    }
  ]
}

Fraud score values: NORMAL, TRUSTED, WARNING, HIGH_RISK
Status values: SUCCESS, SKIPPED, INVALID_INPUT, FAILED

Quality analysis

GET /v2/submission/{submission_id}/quality Returns document image/scan quality assessment. Quality score values: HIGH_QUALITY, LOW_QUALITY, NOT_RELEVANT_DOCUMENT

Document classification

GET /v2/submission/{submission_id}/classification Returns the detected document type. Classification score values: CLASSIFIED, NOT_CLASSIFIED, DOCUMENT_MISSING

Adaptive Decision (customisable layer)

GET /v2/submission/{submission_id}/decision Returns a tenant-specific decision. Available only when enable_decision: true and Adaptive Decision is configured for your tenant by Resistant AI team.

Query parameters

Parameter	Example	Description
`embed`	`fraud`	Embeds the `/fraud` result (e.g., indicators). For full fraud payload (including metadata), fetch `/fraud` directly.

Content extraction (limited access)

GET /v2/submission/{submission_id}/content Returns extracted text/structured content. This endpoint applies only to selected document types from Czechia.

Downloadable PDF report

GET /v2/submission/{submission_id}/report Returns a download URL for a human-readable PDF fraud analysis report.

Query parameters

Parameter	Type	Description
`include_decision`	boolean	Include Adaptive Decision result in the report.

Returns 409 Conflict if fraud analysis or adaptive decision was not completed successfully or was skipped.

Feedback

PUT /v2/submission/{submission_id}/feedback
Stores analyst feedback for the submission. GET /v2/submission/{submission_id}/feedback
Retrieves previously stored feedback.

All custom feedback must be reviewed before submission to ensure it does not contain PII.

Delete a submission

DELETE /v2/submission/{submission_id} Permanently deletes the submission and associated data before the configured retention period.

Returns 409 Conflict if the submission is still being processed.

Common HTTP error codes

Code	Meaning	Recommended action
`400`	Bad request	Validate request body / parameters
`401`	Unauthorized	Verify token is valid and not expired
`403`	Forbidden	Verify scopes/permissions for the tenant/app
`404`	Not found / not ready	Retry with backoff; verify stage/cell + submission_id if persistent
`409`	Conflict	Retry later (e.g., delete while processing)
`415`	Unsupported media type	Ensure correct `Content-Type`
`429`	Too many requests	Back off + retry (exponential backoff + jitter)

For retry and timeout guidance, see Limits & quotas and Polling for results.

Getting started

Integrate

Receiving results

View results in Resistant UI

Operations & Support

Core integration mechanics

Base URL

Step 1 — Create a submission

Request headers

Request body (optional)

Pipeline configuration values

Response — 200 OK

Step 2 — Upload the document bytes

Upload request

Upload rules

Step 3 — Submit submission characteristics (optional)

Response — 204 No Content

Step 4 — Fetch analysis results

Polling semantics (important)

Fraud analysis

Query parameters

Response (example)

Quality analysis

Document classification

Adaptive Decision (customisable layer)

Query parameters

Content extraction (limited access)

Downloadable PDF report

Query parameters

Feedback

Delete a submission

Common HTTP error codes

​Base URL

​Step 1 — Create a submission

​Request headers

​Request body (optional)

​Pipeline configuration values

​Response — 200 OK

​Step 2 — Upload the document bytes

​Upload request

​Upload rules

​Step 3 — Submit submission characteristics (optional)

​Response — 204 No Content

​Step 4 — Fetch analysis results

​Polling semantics (important)

​Fraud analysis

​Query parameters

​Response (example)

​Quality analysis

​Document classification

​Adaptive Decision (customisable layer)

​Query parameters

​Content extraction (limited access)

​Downloadable PDF report

​Query parameters

​Feedback

​Delete a submission

​Common HTTP error codes

Base URL

Step 1 — Create a submission

Request headers

Request body (optional)

Pipeline configuration values

Response — 200 OK

Step 2 — Upload the document bytes

Upload request

Upload rules

Step 3 — Submit submission characteristics (optional)

Response — 204 No Content

Step 4 — Fetch analysis results

Polling semantics (important)

Fraud analysis

Query parameters

Response (example)

Quality analysis

Document classification

Adaptive Decision (customisable layer)

Query parameters

Content extraction (limited access)

Downloadable PDF report

Query parameters

Feedback

Delete a submission

Common HTTP error codes