At Siply, our promise is simple: No ads. No BS. Just water. Every decision we make is driven by that idea — stripping away the gimmicks and giving you the most straightforward way to stay hydrated. That’s what led us to create something a little ambitious: an image recognition feature that estimates how much water is in your cup just by looking at it.
At the heart of this new scan feature is our first-generation machine learning model, Nova, short for NVA-1 (Nova Volume Approximation — Version 1). Nova is an image regression model, trained entirely in-house, to estimate the fluid capacity of common containers. It’s not perfect (yet), but it’s an exciting leap forward for the future of water tracking.
⸻
How Nova Works
When you scan a container using Siply’s scan feature, the app captures a photo and runs it through Nova — a CoreML-powered image regression model. It’s designed to recognize the shape, size, and context of what it sees and predict how many ounces the container can hold.
Nova has been trained using a dataset of real containers — everything from generic water bottles and mugs to popular brands like Hydro Flask and Stanley. The more data we feed it over time, the more accurate it becomes.
Right now, it’s good — but not perfect. Lighting, camera angles, obscure containers, or custom tumblers may throw it off. That’s why when you see a scan result, you can still choose to confirm or adjust the value. Siply puts the control in your hands, and Nova just makes the process faster.
⸻
Constantly Learning (But Only If You Let It)
One thing we’re proud of: Siply never collects your images without permission. Privacy matters. If you enable data sharing in the app settings, your scans may be anonymously and securely shared with us to improve future versions of Nova. If you don’t? No problem — everything stays on your device.
We never store personal data, location info, or anything else tied to you. Shared images are stripped of any metadata and used solely to improve the model’s performance. You’re in full control.
⸻
What’s Coming Next
Nova is version 1 for a reason. In the near future, we plan to:
• Improve accuracy across all lighting and angle conditions
• Expand support to include more container shapes and sizes
• Enable real-time scanning using augmented reality (AR)
• Let users submit new containers to train the model
• Offer manual labeling to improve model precision
Your feedback matters. Every time you use the scan feature and adjust the result, you help inform how we build future models. NVA-2 and beyond will be smarter, faster, and more reliable — and it starts with real-world use.
⸻
Why We Built It
Logging your water shouldn’t feel like a chore. If pulling out your phone and typing in “16oz” is the thing stopping you from tracking your hydration, we wanted to remove that friction. The goal of Nova isn’t to be fancy — it’s to make things effortless. A two-second scan should be all it takes.
That’s the future we’re building at Siply. One where tracking your hydration is something you barely think about because the tools just work.
⸻
Closing Thoughts
Nova is still in development. It’s not flawless — and we’re not pretending it is. But it represents the type of tools we want to build: thoughtful, private, and genuinely useful. We’re excited to share it early and grow it together with you.
If you’re curious to dive deeper into how it works, check out our Learn More page. And as always, read our Terms of Service for how we handle data with care.
Thanks for trusting us with your hydration journey.
— The Siply Team
