Protected by MagenTrust
    MAGEN
    MAGEN Docs

    API Reference

    Complete reference for the MagenTrust public API endpoints.

    Base URL

    https://api.magentrust.ai

    Authentication

    All API requests require authentication via request headers. Include these headers with every request:

    x-magen-key:
    <your-api-key>required
    x-magen-site:
    <your-site-id>required
    Content-Type:application/json

    API keys are rate-limited and domain-restricted. Generate keys in the dashboard.

    Endpoints

    POST/v1/verify

    Verify a session and receive verification context. Returns privacy-safe session information without exposing internal scoring logic.

    Request Headers

    x-magen-key:<api-key>required
    x-magen-site:<site-id>required
    Content-Type:application/json

    Request Body

    {
      "session_id": "sess_abc123def456"
    }

    Response

    {
      "session_id": "sess_abc123def456",
      "verdict": "verified",
      "score": 0.92,
      "risk_band": "low",
      "is_human": true,
      "sdk_version": "1.2.0"
    }
    GET/v1/challenge

    Request an optional challenge for additional verification. Challenges are passive and do not interrupt user experience.

    Request Headers

    x-magen-key:<api-key>required
    x-magen-site:<site-id>required

    Response

    {
      "challenge_id": "ch_xyz789",
      "type": "passive",
      "expires_at": "2024-01-15T10:30:00Z"
    }

    Challenges are optional and used for additional confidence in high-value flows.

    Response Fields

    Verification responses provide context, not internal detection logic. Use these fields for attribution and decision-making.

    FieldTypeDescription
    session_idstringUnique identifier for the verification session
    verdictstringverified | unverified | review
    scorenumberRelative confidence indicator (0-1)
    risk_bandstringlow | medium | high
    is_humanbooleanDerived boolean indicating human confidence
    sdk_versionstringVersion of the SDK that collected signals

    Note: MagenTrust does not expose raw signals or scoring logic. Responses are designed for attribution and decision-making, not detection transparency.