Skip to main content

Overview

TrueParser returns standard HTTP status codes together with JSON error bodies for most API failures. The most common error payload shape is:
{
  "error": "error_code",
  "message": "Human-readable explanation"
}
Some endpoints may also return route-specific fields such as detail.

Authentication Errors

These errors happen when a request is missing a valid access token or the token does not satisfy API policy.
HTTP StatusError CodeMeaning
401missing_tokenNo bearer token was provided.
401token_expiredThe access token has expired.
401invalid_tokenThe token could not be accepted.
403invalid_scopeThe token does not have the required API scope.
403domain_not_allowedThe request origin is not allowed for this application.
503auth_unavailableAuthentication infrastructure was temporarily unavailable.

Document Submission Errors

These errors happen during document ingestion.
HTTP StatusError CodeMeaning
400invalid_inputThe uploaded file is missing or empty.
400document_type_requiredZIP uploads must include an explicit documentType.
400unsupported_formatThe file format is not supported.
400csv_route_requiredCSV uploads must include csvRoute.
400sql_dialect_requiredSQL uploads must include sqlDialect.
400pdf_mode_requiredPDF uploads must include pdfMode.
400invalid_metadatacustomMetadataJson is not valid JSON.
403advanced_extraction_not_allowedThe current plan does not allow advanced PDF extraction.
403license_restrictionThe requested operation is not allowed under the current plan or format entitlement.
413document_too_largeThe upload exceeds the maximum allowed size for the plan.
429rate_limit_exceededThe application exceeded its API traffic limit.
429quota_exceededThe application exceeded a quota window and cannot accept more work right now.
503none fixedThe node is temporarily overloaded.

Result and Status Errors

These errors happen when checking status or retrieving parsed results.
HTTP StatusError CodeMeaning
404not_foundNo job exists for the requested document id.
404result_not_foundThe job completed but no result reference was available.
404document_expiredThe parsed result is no longer available.
422extraction_failedThe parser failed to process the document.
202none fixedThe job is still processing and the result is not ready yet.

Health and Overload Responses

TrueParser also exposes health and overload signals:
EndpointHTTP StatusMeaning
/health/live200The service process is alive.
/health/ready503A required dependency is not ready.
/health/node-state503The node is overloaded and should not receive new traffic.
When a node is overloaded, the API may return a response like:
{
  "status": 503,
  "state": "overloaded",
  "reason": "cpu",
  "retryAfter": 60
}

Quota and Processing Failures

Background processing failures may eventually surface through status polling or result retrieval. Common examples include:
  • quota_exceeded
  • document_quota_exceeded
  • job_timeout
  • processing_failed
  • transient_infrastructure_failure

What Clients Should Do

  • treat the HTTP status code as the primary signal
  • inspect the error field when present
  • use retry with backoff for 429 and appropriate 503 responses
  • poll status for asynchronous jobs instead of assuming immediate completion
Last modified on April 1, 2026