🧠 Here’s What Happens Behind the Scenes:
1. Text Extraction (OCR)
When you upload a screenshot, SmishAlert uses optical character recognition (OCR) to extract the text content — including any phone numbers, links, and phrases.
2. Threat Pattern Matching
The extracted text is then evaluated against a constantly updated database of phishing and social engineering patterns. This includes:
Known phishing phrases and impersonation tactics
Urgent call-to-actions used in scams (e.g., "verify now", "urgent payroll issue")
Formatting signals common in smishing attacks
💼 For business plans, SmishAlert also leverages known context about your organization — including brand terms, executive titles, and communication norms — to customize threat detection to your environment. This helps catch targeted impersonation attempts that generic filters often miss.
3. URL Threat Intelligence
If a link is present, SmishAlert checks it against a globally maintained threat database of suspicious and malicious URLs. We look for:
Known phishing domains
Active malware hosts
Masked redirect URLs (e.g. bit.ly or tinyurl)
4. Contextual AI Scoring
Our AI also analyzes the structure and tone of the message to detect:
Impersonation attempts
Urgency or pressure tactics
Suspicious formatting
This step allows us to flag threats even when no link is present.
5. Final Threat Assessment
Based on all inputs, SmishAlert assigns a threat level — High, Medium, or Low — and delivers the result to your app or iMessage extension instantly.
🔒 Privacy First
All analysis is done securely. Your screenshots and extracted content are encrypted and never used for marketing or third-party access. We only use aggregate insights to improve detection over time.