EMO-Bot | Devpost

GIF
Producing itself
GIF
Face mesh, gesture
GIF
Blinking
Emblem
Web cam gripper
Phone holder
Team
Pixel art

Inspiration

We’ve always been inspired by those lovable robots from movies—Baymax, R2-D2, WALL-E—the ones that aren’t just machines, but companions with real personality. We wanted to bring a bit of that magic into real life. Everyone deserves a buddy who’s there for them, even if it’s made of plastic and bolts. So we set out to build something that doesn’t just do tasks, but feels alive. A little quirky, a little emotional, and a lot of heart.

What it does

EMO-Bot is your personal emotional companion. Whether you’re chilling alone at home or going through a rough day, it’s there for you—no judgment, just vibes. Using real-time emotion detection, it senses your mood and reacts accordingly. If you’re happy, it’ll roll up and celebrate with you. If you’re feeling down, it’ll gently approach and try to comfort you with its little gestures and presence. It doesn’t just recognize how you feel—it responds like a true friend would.

How we built it

We started with a dream and a bag of hex bolts. Then came the glue, duct tape, and a 3D printer—we built the hardware with pure determination (and probably too much Red Bull). EMO-Bot was truly a labor of love… and caffeine.

On the tech side:

🧠 MediaPipe – Used Google’s ML framework for face detection, face mesh, and hand tracking to understand the user’s emotional state.

🎥 OpenCV – Captured and processed real-time video input for emotion recognition.

🔧 mBot Firmware Hacking – We reverse-engineered and modified mBot’s firmware so it could work with our Linux system and compile C++ code smoothly.

🍓 Raspberry Pi – Acted as the robot’s brain, running all the processing and keeping everything on track.

It was duct tape meets deep learning—and somehow, it worked.

Challenges we ran into

Let’s just say… hardware is hard.

We ran into all sorts of issues trying to get everything working together. The Raspberry Pi 5 turned out to be a bit of a power-hungry monster, and powering it efficiently was a constant struggle. Building a robot as software engineers felt like doing surgery with a wrench—we had to figure out balance, structure, how the bot should look, and how to hold it all together (spoiler: duct tape helped, but only so much).

One of the biggest challenges was hacking around the mBot’s closed-source Arduino firmware. Instead of using their existing tools, we scrapped it and built our own custom firmware from scratch to make it work with Linux and C++.

It was chaotic, messy, and held together by sheer willpower (and screws we kept losing).

Accomplishments that we're proud of

EMO-Bot can accurately detect human emotions and gestures using just a camera and MediaPipe, and it actually responds in a way that feels meaningful. Whether you’re smiling or frowning, it vibes with you and reacts in kind. We even gave it a cute little robot voice that makes it feel more alive.

But the biggest win? We hit our original goal: build a robot that could feel like a friend. And we did that. That’s something we’re genuinely proud of.

What we learned

Hardware is hard—but super rewarding when it finally works.

Emotion detection isn’t just about fancy models; it’s about tuning it to real-world, messy human behavior.

Duct tape can fix many things… but not everything.

Building something that “feels alive” is more about personality than perfection.

Bringing lipstick to a hackathon is weirdly useful for facial detection calibration.

Red Bull is not a substitute for sleep (but it does give you wings… for a few hours).

Collaboration, hacking things together, and laughing through failure is what makes hackathons awesome. Use whatever you have with you, it works!!

What's next for EMO-Bot

Through this project, we realized there’s a real need out there—for people who feel lonely, who can’t afford a pet, or simply don’t have the time to take care of one. EMO-Bot could be the perfect alternative: a low-maintenance companion that still brings emotional warmth and connection.

Our next goal is to take EMO-Bot beyond the hackathon and turn it into a real product. We want to refine its design, make it more responsive, and accessible to anyone who needs a little emotional support in their day-to-day life. Whether it’s for kids, the elderly, or busy young adults, we believe EMO-Bot can be that little friend that makes life feel just a bit less lonely.