Gmail AI Threat Scanner: An Open‑Source Security Layer for Your Inbox

Email serves as the primary entry point for cyberattacks, yet most individuals depend solely on whatever built-in filtering their email provider provides. Inspired by the need for a more transparent, customizable, and AI-driven threat analysis, I developed Gmail AI Threat Scanner. This project connects to a user's Gmail inbox, analyzes messages using AI-powered threat detection, and flags suspicious content—all while running locally, ensuring complete control over the process.

What the tool does:

- Scans Gmail messages using AI models for indicators of phishing, fraud, and social engineering

- Highlights risky emails with clear explanations

- Operates entirely client-side using OAuth (no passwords, no server storage)

- Provides a simple, readable interface for reviewing threats

The motivation behind its creation was a desire for a lightweight, open-source tool that offers users visibility into potential threats without the need for enterprise-grade infrastructure. The belief is that security should not be restricted behind corporate paywalls, especially when AI can help individuals defend themselves.

Looking ahead, the project plans to add:

- Multi-model ensemble scoring for more robust threat assessment

- Attachment scanning to broaden protection

- Learning from false positives to improve accuracy over time

The GitHub repository for Gmail AI Threat Scanner can be found at https://github.com/ajaynyayapathi/Gmail-AI-Threat-Scanner. Feedback, issues, or pull requests are welcome.