Relod
ESP32 sensors + consumption forecasting for food businesses. Know when to reorder before you run out. Currently piloting with San Francisco/Berkeley cafés and bars. The physics are simple (distance sensors); the behavioral patterns are not. Accepted to Founders Inc.
Technology Stack
Links
Frequently Asked Questions
How does Relod actually work?
Relod uses time-of-flight sensors to measure container levels, plus humidity and temperature sensors. The system automatically calibrates based on your usage—if you fill to 80% and that's your maximum, it becomes your 100% reference. Edge functions in Supabase calculate optimized metrics. The container learns your consumption patterns: if you typically go from 80% down to 15% before refilling, it will predict when you'll need to reorder. For example, at 23% with a typical 5% usage rate, it knows you'll likely refill after the next use at 18%. The system always offers manual overrides—humans have the last word—while being proactive about helping you stay stocked.
What was the biggest technical challenge in building Relod?
The hardware was the biggest challenge—organizing components, designing custom boards, 3D-printing lids, and learning from mistakes. We started with coffee, using sealed containers, but coffee emits CO2 and caused humidity issues. We learned to add small vents, which accelerated development. We pivoted from B2C coffee consumers to B2B food businesses when cafés showed interest. None of us are TypeScript developers—Carlos is a mechanical engineer at a quantum computing company, I'm a data engineer. We built a full Next.js app with APIs using AI agents like Claude Code, which was an interesting challenge. Forecasting is natural for me as a data person, while Carlos handles the hardware expertly.
Who is Relod best suited for?
Relod is intelligent inventory automation for modern operations—anyone who does inventory. Our main customers are bars and cafés, but it's for all food operations: restaurants, commissary kitchens, hospitality businesses. It provides real-time inventory intelligence for anyone who needs to know how much they have left in their kitchen and cannot afford to run out or have things go bad.
What advice would you give to someone building an IoT product?
Start small and validate everything early. At Founders Inc, we got excited about switching to B2B and I started building a complex multi-tenant, multi-location platform imagining clients like Berkeley Bowl with multiple locations and departments—before we had them as actual clients. It's easy to imagine people will want features when you're building something you care about. Wait for customers to show you—not just tell you—that they need it. Don't build for imaginary scale; build for real needs.
What surprised you most from piloting with cafés and bars?
When we started helping coffee consumers never run out, I thought it was silly and that businesses would be more interested—restaurants actually cannot run out of things. The surprise was how enthusiastic restaurant and café owners were before we even finished explaining what Relod did. They had ideas about features but ultimately said, "I mostly just need to know how much I have left, and I can handle the rest." They wanted notifications when supplies were low and easy ordering in the app, rather than full automation. I don't expect it to stay this way—as people get comfortable, it will take over more of their workflows. But it's incremental, not a jump from 0 to 100.