Red Flag AI Pro API

Request body

{
  "title": "My Sales Page",        // optional, string
  "content": "Your copy here..."   // required, string, min 20 chars
}

Example request

curl -X POST https://redflagaipro.com/api/v1/scan \
  -H "Authorization: Bearer rfp_your_key" \
  -H "Content-Type: application/json" \
  -d '{
    "title": "Landing Page",
    "content": "Earn six figures from home with our guaranteed system..."
  }'

Response

{
  "scan_id": "uuid",
  "title": "Landing Page",
  "score": 60,
  "risk": "medium",
  "flag_count": 2,
  "flags": [
    {
      "category": "income_claim",
      "severity": "high",
      "text_excerpt": "…six figures from home…",
      "description": "Contains an income claim...",
      "suggestion": "Add a clear earnings disclaimer..."
    }
  ],
  "scanned_at": "2026-05-29T11:00:00.000Z"
}

Request body

{
  "title": "Auto-published blog draft", // optional, string
  "content": "The text to evaluate...", // required, string, min 20 chars
  "threshold": 50                       // optional, 0-100, default 50
}

Example request

curl -X POST https://redflagaipro.com/api/v1/enforce \
  -H "Authorization: Bearer rfp_your_key" \
  -H "Content-Type: application/json" \
  -d '{
    "title": "Auto-published blog draft",
    "content": "Earn six figures from home with our guaranteed system...",
    "threshold": 50
  }'

Response

{
  "decision_id": "uuid",
  "allowed": false,
  "valid_until": "2026-08-24T11:01:00.000Z",
  "score": 35,
  "threshold": 50,
  "risk": "high",
  "flag_count": 2,
  "flags": [ { "category": "income_claim", "severity": "high", "..." } ],
  "verify_url": "https://redflagaipro.com/verify?id=...",
  "governing_record": { "id": "uuid", "decision": "...", "owner_name": "..." },
  "checked_at": "2026-08-07T11:00:00.000Z"
}

A blocked decision (allowed: false) is sealed with an independent timestamp and a public verify link. Allowed decisions are stored but not individually sealed, to avoid flooding the chain with routine checks. governing_record is the current boundary authorization record for the API key making this call, if one exists and has not expired — null if nothing currently governs this key. Named on the decision itself, not left as a separate record you have to go match up by hand.

valid_until is how long this specific answer can be relied on, not how long the key is authorized in general. An ALLOW is a stateless answer to one question asked at one moment. Nothing stops a caller checking once and then dispatching several actions off the back of that single answer, and a falsifier that fires a second later cannot reach back and recall anything already launched under it. Set to 60 seconds from the check, or the governing record's own expiry if that comes sooner, whichever is closer. Treat an ALLOW past this timestamp as stale and check again rather than assuming it still holds. null when the decision is blocked, since a block has nothing to expire.

Response

{
  "bundle": {
    "version": 1,
    "decision_id": "uuid",
    "checked_at": "2026-08-10T11:00:00.000Z",
    "title": "Auto-published blog draft",
    "score": 35,
    "threshold": 50,
    "allowed": false,
    "block_reason": "content_score",
    "flag_count": 2,
    "flags": [ { "category": "income_claim", "severity": "high" } ],
    "governing_record": {
      "id": "uuid",
      "decision": "...",
      "owner_name": "...",
      "owner_role": "...",
      "authority_mode": "human_decides",
      "expires_at": "2026-12-01T00:00:00.000Z",
      "permission_fingerprint": "pf-a1b2c3d4",
      "fingerprint_intact_at_export": true
    },
    "exported_at": "2026-08-10T11:05:00.000Z",
    "exported_by": "redflagaipro.com"
  },
  "signature": "base64...",
  "algorithm": "ed25519",
  "public_key_pem": "-----BEGIN PUBLIC KEY-----...",
  "public_key_url": "https://www.redflagaipro.com/api/verify/signing-key"
}

Verify offline with the standalone script at /verify-decision-bundle.js, no dependencies beyond Node's built in crypto module: node verify-decision-bundle.js bundle.json. Save the public key alongside your bundles rather than re-fetching it each time — the whole point is not having to trust the server again at the moment you check.

Query parameters

limit   integer   Max results to return (default 20, max 100)
offset  integer   Number of results to skip (default 0)

Example request

curl https://redflagaipro.com/api/v1/scans?limit=10 \
  -H "Authorization: Bearer rfp_your_key"

Response

{
  "scans": [
    {
      "id": "uuid",
      "title": "Landing Page",
      "score": 60,
      "status": "complete",
      "created_at": "2026-05-29T11:00:00.000Z"
    }
  ],
  "total": 42,
  "limit": 10,
  "offset": 0
}

Payload

{
  "event": "scan.completed",
  "scan_id": "uuid",
  "title": "Landing Page",
  "score": 60,
  "flag_count": 2,
  "flags": [
    {
      "category": "income_claim",
      "severity": "high",
      "suggestion": "Add a clear earnings disclaimer..."
    }
  ],
  "scanned_at": "2026-05-29T11:00:00.000Z"
}