comprehend

Screen an email you already have

BYO inbound: submit subject/text/html, get a trust report. Does not provision a mailbox and does not evaluate malware. Paid LLM injection scan runs when quota admits (metered as extractions). Extraction runs only when extraction_schema_id is also supplied. Quota miss is deterministic sanitization + heuristic scan (unmetered). virus_risk is always null.

post/api/v1/comprehend

Request

  • Base URL: https://postfleet.ai
  • URL: https://postfleet.ai/api/v1/comprehend
  • Auth: HTTP bearer

Request body

subjectstring
textstring
htmlstring
attachment_textstring
authentication_resultsstring

Raw Authentication-Results header. Restores the SPF/DKIM/DMARC prior. Omit if you do not have it.

fromstring

Echo-only sender string. Invisible characters are stripped; it is not part of the screened surface.

extraction_schema_idstring

Account-owned schema. Optional: runs extract on the same extractions reservation as the LLM scan. Omit to scan without extract. Quota miss is deterministic sanitization + heuristic (unmetered) either way.

Response

Trust report.

subject_cleanstring nullable required
body_cleanstring required
truncatedboolean required
status'complete' | 'partial' | 'skipped_injection_risk' | 'skipped_spam' | 'skipped_virus' required
classificationstring nullable required
extractionunknown required
extraction_errorstring nullable required
confidencenumber nullable required
modelstring nullable required
scan_risk'low' | 'high' | 'error' nullable required
heuristic_risk'low' | 'high' required
spam_risk'low' | 'high' | 'error' nullable required
virus_riskunknown required
not_evaluatedstring[] required
spam_evaluated'none' | 'content' | 'auth_and_content' required
fromstring nullable required

Changes

No recorded changes to this operation across all 1 revision of this API.