Countr

Stockouts and overstock are killing your margins. Countr gives you real-time inventory visibility so you always know what you have, where it is, and when to reorder.

Comment

Inspiration

While biking through the Vietnamese countryside, our team stumbled upon something that stuck with us: local shop owners visibly struggling to locate items when customers asked for them. Drinks, snacks, everyday goods — things that should be easy to find were causing real friction. It wasn't a lack of effort; it was a lack of visibility into what they actually had. That moment made it clear: small, independent retailers are underserved by modern inventory tools, which are often too expensive, too complex, or too manual. We built Countr to change that.

What it does

Countr is a computer vision–powered inventory management system designed for small and local businesses. Point a camera at your shelves, and Countr automatically counts your stock, tracks levels in real-time, and flags items running low before they run out. An analytics dashboard gives shop owners a clear picture of their inventory at a glance — no spreadsheets, no manual counting, no guesswork.

How we built it

Countr is built around a computer vision model that detects and counts physical items from a camera feed. We trained and integrated the CV pipeline to handle real-world shelf conditions — varying lighting, cluttered arrangements, and diverse product shapes. On top of the CV core, we built a real-time tracking layer, a low-stock alerting system, and an analytics dashboard to surface actionable insights for store owners.

Challenges we ran into

Our biggest challenge was scope: building a full, functional system — from CV inference to a working dashboard — within the hackathon window. We had to make fast decisions about what to cut, what to simplify, and where to focus our energy. Coordinating across the stack under time pressure pushed us to work lean and communicate constantly.

Accomplishments that we're proud of

We're proud that Countr actually works end-to-end. From a live camera feed to real-time stock counts to dashboard alerts — the pipeline holds together. More than the tech, we're proud that we built something rooted in a real, observed problem rather than a hypothetical one. The local shop owners we thought about during that bike ride are exactly the people this is for.

What we learned

We learned how quickly a well-scoped idea can come together when the team is aligned. Working under constraints forced us to prioritize ruthlessly and ship fast. We also deepened our understanding of applying computer vision to practical, physical-world problems — and how much detail goes into making CV reliable outside of a controlled environment.

What's next for Countr

We want to bring Countr to the local businesses that inspired it. Next steps include expanding the CV model's accuracy across a wider range of products, adding a mobile-friendly interface so shop owners can check inventory from their phone, and exploring a lightweight hardware setup (like a Raspberry Pi + camera) for affordable, always-on shelf monitoring. Long-term, we'd love to integrate with ordering systems so Countr doesn't just alert you when stock is low — it reorders for you automatically.

Built With

Share this project:

Updates