Skip to main content
Beyond the core fraud analysis workflow, the Documents API includes optional capabilities that may be enabled per submission (via flags) or per tenant (by configuration).
If you’re looking for the end-to-end flow (create submission → upload → fetch results), see Core Integration Mechanics or start with Quickstart API.

​
Highlights

Adaptive Decision

Tenant-specific decision output based on analysis inputs (optional add-on).

Downloadable PDF report

Generate a downloadable human-readable report (optional).

​
Decisioning and outputs

​
Adaptive Decision (optional)

Adaptive Decision returns a tenant-specific decision derived from analysis inputs (e.g., fraud, and optionally other configured inputs).
  • Enable: set enable_decision: true when creating the submission
  • Fetch: GET /v2/submission/{submission_id}/decision
See: Adaptive Decision

​
Classification (optional / tenant-config dependent)

Classification identifies the document type.
  • Run standalone: pipeline_configuration: "CLASSIFICATION_ONLY"
  • Fetch: GET /v2/submission/{submission_id}/classification
Availability and behavior may depend on tenant configuration.

​
Quality (optional / tenant-config dependent)

Quality assesses scan/photo quality.
  • Run standalone: pipeline_configuration: "QUALITY_ONLY"
  • Fetch: GET /v2/submission/{submission_id}/quality
Availability and behavior may depend on tenant configuration.

​
Analyst workflows

​
Analyst feedback

Stores analyst feedback for a submission.
  • Submit: PUT /v2/submission/{submission_id}/feedback
  • Retrieve: GET /v2/submission/{submission_id}/feedback
Do not include PII in feedback or comments. Feedback may be retained to support product improvement.
See API Reference for the current request schema.

​
Exports

​
Downloadable PDF report

Generates a download URL for a PDF fraud analysis report.
  • Fetch: GET /v2/submission/{submission_id}/report
  • Optional: include_decision=true
See: Downloadable PDF report

​
Limited access endpoints

​
Content extraction (limited — Czech bank statements only)

The /content endpoint is not available by default and is currently offered only for selected Czech documents (bank statements). It is not a general OCR feature:
  • We do not run OCR across arbitrary PDFs/images.
  • We do not return transcribed/free-form document text for other document types.
  • Fetch: GET /v2/submission/{submission_id}/content
If you believe your use case qualifies (Czech bank statements), contact Resistant AI to discuss enablement for your tenant.

​
Quick reference

CapabilityHow to enableEndpoint
Adaptive Decisionenable_decision: trueGET /v2/submission/{id}/decision
Classificationpipeline_configuration: "CLASSIFICATION_ONLY"GET /v2/submission/{id}/classification
Qualitypipeline_configuration: "QUALITY_ONLY"GET /v2/submission/{id}/quality
Feedback—PUT/GET /v2/submission/{id}/feedback
PDF report—GET /v2/submission/{id}/report
Content extraction (limited — CZ bank statements)tenant configurationGET /v2/submission/{id}/content
For status meanings (SUCCESS, FAILED, INVALID_INPUT, SKIPPED) and polling guidance, see Concepts and Polling for results.