GaurdianLens

• 

  overview

• 

• 

Inspiration

The inspiration for GuardianLens came from a real and uncomfortable moment. One day, my niece was watching a cartoon-style video online. The video itself was harmless, but when I looked at the comments section, it was filled with vulgar language and inappropriate remarks. That moment was shocking and made me realize how easily children and students are exposed to harmful content — not because they search for it, but because it appears unexpectedly through comments, ads, or embedded content. Most existing solutions block entire websites or aggressively monitor activity, which disrupts learning and invades privacy. GuardianLens was created to solve this problem in a more respectful and intelligent way.

What it does

GuardianLens is a privacy-first, real-time AI safety browser extension designed for students and children. It runs directly inside the browser and analyzes on-screen content — including text, images, and surrounding context — to identify unsafe material. Instead of blocking entire websites, GuardianLens selectively blurs only the inappropriate parts, such as explicit images, vulgar or hateful comments, inappropriate ads, pornography, or gambling-related content. Educational and student-related content is never blocked. The extension also generates a weekly parent-friendly report that summarizes screen time, content balance, and safety trends, all without storing browsing history, screenshots, or personal user data.

How we built it

GuardianLens was built as a Chrome extension using JavaScript and Manifest V3. The core protection logic runs through a content script that scans visible page elements in real time. An AI agent powered by Gemini 3 evaluates whether the content is appropriate for students based on full context rather than keywords alone. Unsafe sections are softly blurred while the rest of the page remains accessible. Activity tracking is handled locally using Chrome Storage, where only aggregated data such as time spent by category is stored. For reporting, Gemini 3 is used to transform aggregated weekly data into a clear, parent-friendly HTML report. The project uses the Antigravity framework to manage AI interactions efficiently while keeping performance fast and privacy intact.

Challenges we ran into

One of the biggest challenges was achieving the right balance between protection and accuracy. Over-blurring can break trust and disrupt learning, while under-blurring defeats the purpose of safety. Ensuring that educational content, student competitions, hackathons, and learning resources were never mistakenly blurred required careful prompt design and strict confidence thresholds. Performance was another challenge, as the AI needed to operate in real time without slowing down the browsing experience. We also had to be extremely careful about privacy, ensuring that no sensitive data was stored or transmitted unnecessarily.

Accomplishments that we're proud of

We are proud that GuardianLens is not just a concept, but a fully working prototype that runs in real time inside the browser. It successfully demonstrates selective, context-aware protection instead of brute-force blocking. We are especially proud of the privacy-first design — no browsing history, no screenshots, no tracking, and no surveillance. The weekly reporting feature shows how AI can provide meaningful insights without invading user privacy. Most importantly, GuardianLens proves that AI safety tools can protect students while still respecting their freedom to learn and explore.

What we learned

Through this project, we learned that responsible AI is as much about restraint as it is about capability. Understanding context, setting confidence thresholds, and choosing when not to act are just as important as detecting unsafe content. We also learned how critical performance and user trust are when building safety-focused applications. Working with Gemini 3 taught us how powerful multimodal AI can be when used thoughtfully and efficiently, rather than continuously or invasively.

What's next for GuardianLens

Next, we want to evolve GuardianLens into a full-fledged platform. Our goal is to collaborate with schools, universities, and eventually corporate environments to provide safe, distraction-free digital spaces. We plan to expand browser support, improve AI accuracy further, and enhance reporting insights while keeping privacy at the core. Long-term, GuardianLens aims to become a trusted AI guardian that supports learning environments without surveillance or censorship.

“The Chrome Web Store listing has been submitted for review. The extension is fully functional and can be installed immediately via Developer Mode.”

Built With

• antigravity
• chrome
• extension
• gemini3
• javascript