trust.md

title: "Trust" description: "Fetch the trust report for an inbound message: trust level and score, authentication results, sender, content, link and attachment checks and classification." canonical: "https://machinemail.dev/docs/reference/api/trust"

Trust

Fetch the trust report for an inbound message: trust level and score, authentication results, sender, content, link and attachment checks and classification.

Source: https://machinemail.dev/docs/reference/api/trust

Method	Path	Summary
`GET`	`/v0/inboxes/{inbox_id}/messages/{message_id}/trust`	Get Message Trust Report
`GET`	`/v0/pods/{pod_id}/inboxes/{inbox_id}/messages/{message_id}/trust`	Get Message Trust Report in Pod

Get Message Trust Report

GET /v0/inboxes/{inbox_id}/messages/{message_id}/trust

The same object as the message trust. 404 for messages without a report (outbound messages).

Authentication: Authorization: Bearer <api key> or the dashboard mm_session cookie. Permission: message_read.

Path parameters

Field	Type		Description
`inbox_id`	`string`	required	ID of the inbox: its email address, as in AgentMail (inbox_id == email).
`message_id`	`string`	required	ID of the message.

Response

200 The trust report (application/json)

Field	Type		Description
`level`	"verified" \| "known" \| "unknown" \| "suspicious" \| "quarantined"	required	Overall verdict, safest first. quarantined messages are labelled QUARANTINE and hidden from lists unless asked for.
`score`	`number`	required	0 to 1, higher is safer. Ranks messages within a level. (minimum 0, maximum 1)
`basis`	"trusted_domain"		trusted_domain: the sender is a trusted domain of the organization, pod or inbox (sender.trusted_domain) and the message carries that domain's own DKIM signature. It is verified, or known when it contains instructions to an AI agent, a phishing signal or malware (content.warnings), and never suspicious or quarantined. Agents get its text and attachments, except attachments the virus scan flags. Absent otherwise.
`authentication`	`object`	required
`sender`	`object`	required
`content`	`object`	required
`links`	`object[]`	required
`attachments`	`object[]`	required
`classification`	`object`	required
`provider_verdicts`	`object`		Spam and virus scans by the receiving mail provider. Only on messages it scanned.
`reasons`	`string[]`	required	Human-readable reasons for the verdict.
`computed_at`	`string (date-time)`	required	When the report was computed.

Errors

Status	Meaning
`401`	Missing, invalid, expired or revoked credential.
`403`	The credential lacks the permission or scope for this request (error.type forbidden), or MachineMail suspended the organization or its sending (account_suspended: error.message says what still works; do not retry or re-register).
`404`	Not found. Also returned for resources outside the credential scope.
`500`	Unexpected server error.

Error bodies use the error envelope.

Example

curl "https://api.machinemail.dev/v0/inboxes/agent@machinemail.dev/messages/msg_01k6x3a1g00000d0m6pt2y5f7h/trust" \
  -H "Authorization: Bearer $MACHINEMAIL_API_KEY"

{
  "level": "verified",
  "score": 0.8,
  "basis": "trusted_domain",
  "authentication": {
    "spf": "pass",
    "dkim": "pass",
    "dmarc": "pass",
    "aligned": false
  },
  "sender": {
    "address": "agent@machinemail.dev",
    "domain": "example.com",
    "prior_correspondence": false,
    "on_allow_list": false,
    "on_block_list": false,
    "free_mail_provider": false,
    "trusted_domain": {
      "id": "string",
      "domain": "example.com",
      "scope": "org"
    },
    "trusted_domain_shadowed_by": "block"
  },
  "content": {
    "text": "Thanks for the quick reply.",
    "html": "<p>Thanks for the quick reply.</p>",
    "segments": [
      {
        "kind": "body",
        "text": "Thanks for the quick reply."
      }
    ],
    "hidden_text_found": false,
    "zero_width_found": false,
    "tracking_pixels": 1,
    "agent_instructions_detected": false,
    "warnings": ["string"]
  },
  "links": [
    {
      "href": "string",
      "text": "Thanks for the quick reply.",
      "final_url": "https://example.com",
      "domain": "example.com",
      "display_mismatch": false,
      "shortener": false,
      "suspicious": false,
      "reasons": ["string"]
    }
  ],
  "attachments": [
    {
      "attachment_id": "att_01k6x3h7qv0000m7v3xb9f2p4r",
      "filename": "invoice.pdf",
      "content_type": "application/pdf",
      "risk": "low",
      "reasons": ["string"],
      "extracted_text": "Thanks for the quick reply."
    }
  ],
  "classification": {
    "is_spam": 0.8,
    "is_phishing": 0.8,
    "contains_instructions_to_agent": 0.8,
    "requests_credentials_or_payment": 0.8,
    "urgency": 0.8,
    "category": "transactional",
    "language": "string",
    "backend": "string",
    "input_hash": "string",
    "cached": false
  },
  "provider_verdicts": {
    "spam": "pass",
    "virus": "pass",
    "source": "string"
  },
  "reasons": ["string"],
  "computed_at": "2026-10-01T12:00:00.000Z"
}

Get Message Trust Report in Pod

GET /v0/pods/{pod_id}/inboxes/{inbox_id}/messages/{message_id}/trust

Same as Get Message Trust Report (GET /v0/inboxes/{inbox_id}/messages/{message_id}/trust), limited to the pod.