ript | Devpost

The Why.

Gyms can be scary. We’ve all felt it: walking in, not knowing how to use a machine, feeling eyes on us, worrying if we’re doing it “right.” Turns out, The Times says 3 out of 4 women avoid gyms for those exact reasons. Spoiler: that includes all four of us. Add in the cost (women-only gyms often start at $50/month) and the fact that working out at home is harder than it looks… and suddenly staying active feels like an uphill battle. We’ve had injuries, we’ve had moments of frustration staring at a YouTube video that doesn’t look anything like our bodies. We wanted something better. Something that feels supportive, safe, and built for real people; not just athletes or influencers.

The What.

That’s where Ript comes in.

Think of Ript as a personal trainer who lives on your laptop (or eventually your phone): It watches your form through your camera and uses AI to spot issues in real time. Instead of yelling, it gives gentle cues like “straighten your back” or “try a little deeper”, the kind of advice you’d get from a supportive friend. It keeps track of your progress: not in pounds or aesthetics, but in consistency, strength, and health. It’s built for everyone: adaptive workouts and body-positive coaching. No judgment. No pressure. Just you, moving at your own pace.

The How.

We split up tasks and dove in: Frontend: React + TypeScript for a clean, modular UI. Context API for auth and workout state. Backend: Node.js + Express, MongoDB Atlas for user data and workout logs. AI: TensorFlow.js and experiments with MediaPipe for real-time pose detection (all in the browser, no fancy equipment needed). Design: Accessibility first. Every feature, from notifications to adaptive cues, came from asking, “Would this make working out feel less intimidating?”

The Ups.

Real-time pose detection in the browser. We cheered when it finally worked. Building a full-stack app with AI in just a few days. Most importantly: creating something that feels like us. As people who’ve felt excluded from fitness spaces, building Ript was a way of building the kind of support we wish we had.

The Downs.

We were first-time users of OpenCV and MediaPipe, and the learning curve was… steep. (Like, “why are there 33 keypoints but none of them are where we want them” steep.) Making the AI fast enough to give feedback while you’re moving was a real balancing act. And writing feedback that feels like encouragement instead of criticism? Way harder than it sounds.

The Takeaways.

AI doesn’t solve problems on its own, you have to design around people’s feelings to make it work. MongoDB is actually a great fit for messy, real-world workout data. And hackathons? They’re just as much about laughing through bugs at 3 a.m. as they are about the final demo.

The Next Steps.

Bringing Ript to mobile (React Native). Adding streaks, badges, and fun ways to stay motivated. Expanding the exercise library (yoga, weights, cardio). Testing with real communities to keep inclusivity at the heart of it.

The Goal.

Ript isn’t about six-packs or PRs. It’s about making fitness feel safe, encouraging, and accessible. We want you to look at your screen and feel celebrated, not judged. So instead of asking, “Do I belong in the gym?”, Ript helps everyone answer, “Yes. Right here. Right now.”

We’re Ript: fitness that flexes for everyone.