Ping-Pill

Inspiration

To stabilize my mental health, I need to take vitamins and medication regularly — but I often forget whether I’ve taken them. I wanted a companion that could notice and remind me, without needing to scan QR codes or press buttons. That’s how the idea for Owl-Pill was born: a visual assistant that sees, remembers, and supports my routine.

What it does

Ping-Pill visually detects if you’ve taken your daily pills or not. It uses a tiny camera hidden in an owl-shaped gadget to watch the pillbox and track changes. The data is stored on a server and you can ask “Did I take my pills today?” — and get an exact answer.

How we built it

We used Raspberry Pi Zero 2 W and Camera Module 3 inside 3D-printed owl body. Backend is Node.js + MySQL on VPS, frontend is Vite + React. LLM runs localy via Mistral 7B and Ollama, Eleven Labs adds voice. Bolt AI helped me to generate/refactor code even if I’m not developer.

Challenges we ran into

• I’m not a programmer, so it took time to understand system architecture and learn how domain, backend and DB work.
• Working with camera and hardware was completely new.
• I printed the shell in PETG, and needed to tweak settings to get clean prints.

Accomplishments that we’re proud of

• I build LLM-based voice interface in 3 hours with Bolt, with no real programming background — I’m a designer.
• MVP version of pill-tracking assistant works and looks like real home gadget, not some hospital tool.

What we learned

• Combining vision detection with LLMs for wellness use
• Hosting models locally for privacy and speed
• Designing tools that are useful but also emotionally supportive

What’s next for Owl-Pill

• Add support for multiple people in same house
• Connect to calendar and notify about missed pills
• Add LED/sound cues to remind softly
• Make more health-animal gadgets

Built With

• 3dprinting
• bolt
• boltai
• cameramodule3
• elevenlabs
• elevenlabsapi
• mistral7b
• mysql
• node.js
• ollama
• petg
• raspberrypizero2w
• react
• vite