AI Shield

Detect scams before they cost you.

AI Shield analyzes suspicious messages and visible screen text using local machine learning, behavioral signals, and security rules—while keeping sensitive content under your control.

Local-first analysis Consent-based processing Explainable risk scoring Optional encrypted logging

Message Protection

Inspect a message before you trust it

Paste suspicious content for a local verdict, then choose whether to verify it with the AI Shield API.

  1. Input Paste an email, SMS, or chat message.
  2. Local check Analysis runs in this browser after you opt in.
  3. API choice Content is sent only when you select verification.
  4. Verdict Review the score, detected reasons, and next steps.

Analysis & storage choices

Allow analysis, then paste a message for a local check.

Local Message Verdict

0

Ready when you are

Allow analysis and paste a message to see a local risk assessment.

Runs in this browser

Detected reasons

    Suspicious patterns

      Recommended actions

        Optional API Verification

        0

        Not Verified

        Use API verification when you want a server-confirmed result.

        Starts only from the verification button

        Detected reasons

          Suspicious patterns

            Recommended actions

              Screen Analysis

              Check suspicious text on your screen

              Paste visible text for a browser-local verdict, or capture a screen for OCR by the locally hosted AI Shield service.

              1. Input Paste text or choose a screen to capture.
              2. Local check Pasted text is analyzed in this browser.
              3. API choice Verification sends the visible text to the API.
              4. Verdict See the score, suspicious patterns, and guidance.

              Allow screen analysis, then paste text or choose a screen to capture.

              Local Screen Verdict

              0

              Ready when you are

              Allow analysis and add visible text to see a browser-local verdict.

              Pasted text runs in this browser

              Detected reasons

                Suspicious patterns

                  Recommended actions

                    Optional API Verification

                    0

                    Not Verified

                    Send the extracted or pasted text when you want an API-confirmed result.

                    Starts only from the verification button

                    Detected reasons

                      Suspicious patterns

                        Recommended actions

                          Architecture

                          How AI Shield makes its decisions

                          1. Multi-layer detection

                          AI Shield combines local machine-learning classification using Naive Bayes, rule-based indicators, and behavioral analysis to identify phishing, impersonation, coercion, and scam patterns.

                          2. Minimal encrypted logging

                          Logging is optional. When enabled for an API request, AI Shield stores encrypted audit metadata, hashed identifiers, and a content preview only if separately selected.

                          3. Explainable decisions

                          Each completed verdict shows detected signals, suspicious patterns, and recommended actions so users can understand the score.

                          4. Ready for extension

                          The modular detection and API layers can support future browser extensions, OCR improvements, voice-scam detection, and mobile workflows.

                          Product Feedback

                          Help improve AI Shield

                          Share a rating and optional product comment. Feedback is separate from message analysis and screen verification.