-
-
Logo
Inspiration Picture this: you’re signing up for something online, and bam—there’s a 10-page privacy policy staring you down. Total gibberish, right? We’ve all just clicked “agree” and crossed our fingers, but deep down, it bugs you. What are they doing with my info? Are they selling it? Then there’s those sketchy websites that feel off—like that too-perfect deal that’s probably a scam. That frustration hit us hard. Some of us had friends who got burned by bad sites, and we’d had our own close calls. So, we thought: what if we could build something to make the internet less of a gamble? That’s where VortexWatch came from—a way to take back control and browse with confidence.
What it does VortexWatch is like your trusty sidekick when you’re surfing the web. It’s a browser extension that kicks in the second you land on a site. First, it digs up the privacy policy (no more hunting for it yourself) and scans it for shady stuff, think sneaky data-sharing lines or vague “we’ll protect you” promises that don’t mean much. It doesn’t just wave a red flag; it shows you the exact sentences that tripped its alarm and slaps a safety score on the site, from 1 to 10. Super simple. But wait, there’s more! While it’s dissecting the policy, it’s also eyeballing the site’s HTML. Weird scripts? Shady pop-ups? It’ll call them out and explain why they’re trouble. Then it labels the site Safe, Suspicious, or Unsafe. If you’re stuck on a dodgy site but need what it’s got, VortexWatch has your back it suggests safer alternatives that do the same thing. It’s like a security guard and a helpful friend all in one.
How we built it We built VortexWatch by combining some cutting-edge AI tools and web scraping tech. For the AI side, we tapped into Cohere, OpenAI, and Gemini to analyze privacy policies and dig into website data. These models helped us break down dense legal jargon and spot anything sketchy that might signal an unsafe site. On the scraping front, we used Beautiful Soup and Selenium to pull apart HTML structures, letting us examine how websites are built and flag any red flags in their design or behavior. The whole operation runs on Google servers, which keeps things stable and ready to scale. We leaned heavily on API calls to tie everything together, making sure all these pieces, AI analysis, web scraping, and server power, worked as a smooth unit right in the browser.
Challenges we ran into One of our biggest headaches was juggling the AI APIs. We started with Cohere, OpenAI, and Gemini, but each had its own quirks. We ended up swapping them out and tweaking things over and over to get the right mix of accuracy and reliability. It was a bit of a grind to keep everything playing nice together. Speed was another major hurdle. Users don’t want to wait around while we analyze a privacy policy or scrape a site, they need answers fast. We spent a ton of time optimizing our code and testing different API setups to shave down processing time without skimping on the depth of our checks. It was a balancing act, but we got there.
Accomplishments that we're proud of We’re really happy with how VortexWatch turned out. The AI can now sift through privacy policies like a pro, picking out suspicious clauses and giving users clear, useful insights. The scraping tools are just as solid, catching weird design patterns that might mean a site’s up to no good. Together, they give users a solid safety net. The interface is something we’re especially proud of, it takes all this complicated analysis and boils it down into a simple safety score (Safe, Suspicious, Unsafe) that anyone can understand. Early feedback has been awesome, with users loving how quick and straightforward it is.
What we learned This project taught us a ton. Wrestling with multiple AI APIs showed us how powerful they can be, but also how much work it takes to get them working in sync. Speed turned out to be just as critical as we thought, users want instant feedback, and we had to rethink a lot of our approach to make that happen. We also learned a lot about building for real people. Turning complex data into something simple and intuitive took a lot of trial and error, and listening to user input was key to nailing it.
What's next for Vortex Watch We’re not done yet! Next up, we’re fine-tuning the AI to catch even trickier policy language and stay ahead of new scam tactics. We’re also growing our list of safe site alternatives, so users have solid options when something’s flagged as risky. Down the road, we’d love to bring VortexWatch to mobile browsers for full-device protection. We’re also thinking about adding customizable alerts, so users can tweak how strict or lenient the extension is based on their own comfort level. There’s a lot more to come!
Log in or sign up for Devpost to join the conversation.