Why I built a fishing log with no account and no cloud

I'm Scott. I live in North Dakota, I work in tech and consulting, and I've been on the water just about as long as I can remember — fly rod, spinning reel, boat, shoreline, prairie reservoirs to mountain streams to Pacific tailraces and just about everywhere in between.

Every time I had a great day on the water, I wanted to know why. What was the pressure doing? Which lake, which hour, which lure? But the tools to answer that never fit the moment. Either I was typing a novel into a notes app while fish were biting, or I was handing my catch history to a social network that wanted an account, a login, and a copy of every spot I'd ever found. It felt like trying to navigate the Missouri with the instruments of Lewis and Clark.

Speed had to come first

A fishing log only works if you actually use it, and you only use it if it's faster than not using it. So the whole app is built around one number: about five seconds from "fish in the hand" to "logged." Snap a photo, the on-device AI names the species, and GPS, weather, moon phase, barometric pressure, and the water body all fill themselves in. No typing, no menus, no homework at the boat ramp.

Privacy wasn't a feature — it was the floor

The second I decided this was my log, the architecture got simpler. No account. No cloud. No analytics SDKs. Your catches, photos, and coordinates live in a local database on your phone, and the only network calls the app makes are anonymous — weather, a water-body name, map tiles. When you share a Brag Card, the GPS is stripped, because the fastest way to lose a honey hole is to post it with a pin on it. If you want the detail, it's all in the privacy-first design and the privacy policy.

The AI had to run in your pocket

The remote lakes where I actually fish are exactly the places a cloud-based "AI fish ID" app gives up. So the model had to live on the device. Elevated Fishing bundles three real neural networks — detection, segmentation, and a 639-species classifier — and runs them locally, with rerankers that know a saltwater grouper shouldn't top the list on a North Dakota lake. It's honest about uncertainty, too: when a photo isn't clear, it says so instead of guessing. Here's how the AI fish ID works.

What it's for

I built the log I wanted: fast enough to use on every fish, private enough that the data is only ever mine, and smart enough that, over a season, the patterns behind my best days start to show up on their own. A bad day of fishing is still a good day — but this makes sure I remember the great ones.

— Scott, founder, Elevated Fishing