Inspiration

---

The internet is a powerful tool, but it comes with risks - children can be accidentally exposed to unsafe content, and adults often lose focus due to constant digital distractions. We wanted to build something simple, effective, and autonomous that protects users without requiring manual intervention.

What it does

---

Censorra is a Chrome extension that automatically detects NSFW content using AI and closes the entire browser window where such content is found. It acts as a background agent that runs silently, ensuring digital safety for children and improved focus for adults.

How we built it

---

• We used Chrome Extension APIs to monitor open browser windows.

• A background service worker classifies content in real time by sending requests to the Hugging Face Inference API.

• If NSFW material is detected, the service worker autonomously triggers the window-closing action.

• Content scripts & CSS provide basic feedback by styling blocked elements during processing.

• The project is structured with TypeScript + Vite for modularity and type safety, and includes React + shadcn/ui setup for possible future expansion into user dashboards or parental controls.

Challenges we ran into

---

• Ensuring accurate classification while keeping performance smooth in the browser.

• Handling browser-level actions like closing windows reliably across different websites.

• Integrating AI inference calls without adding excessive latency to the user’s browsing experience.

• Striking a balance between being strict on unsafe content and avoiding false positives that might frustrate users.

Accomplishments that we're proud of

---

• Building a fully autonomous AI agent that acts in real time without user input.

• Creating a working prototype that demonstrates end-to-end automation - from content detection to decisive action.

• Showing how a simple extension can be expanded into a broader digital well-being tool.

What we learned

---

• How to combine AI inference APIs with Chrome Extension APIs to create automated workflows.

• The importance of background service workers in enabling continuous monitoring and action.

• The practical challenges of applying AI moderation in real-world browsing.

• That even small, focused AI agents can make a significant impact on safety and productivity.

What's next for Censorra

---

• Adding a configurable dashboard for parents and users to set their own safety rules.

• Expanding beyond NSFW detection to include time-wasting sites and distracting content.

• Offering granular actions (like blocking just a tab or replacing images with placeholders) instead of always closing the entire window.

• Improving classification accuracy by experimenting with different AI models.

• Scaling towards a cross-platform solution (desktop + mobile browsers).

Built With

• background-service-worker
• chrome-extension-apis
• content-scripts
• huggingface-inference-api
• react
• tailwindcss
• typescript
• vite