Nature Radar

Inspiration

Experiences in nature are getting popular, with few specialized ways to remember them.

What it does

Elevates your experience with nature. Main functions: Nature journalling through images, infomation of wildlife sightings in your area, and a community tab where you can share your findings.

How we built it

Ideation with Claude code, tinkering with Tensorflow Lite and iNaturalist API, Supabase hosting the images and SQLite to run small AI models on device.

Challenges we ran into

Design decisions regarding app layout to make the experience with the screen minimal for the user.

Accomplishments that we're proud of

Two step visual identification with a generalist AI doing the wildlife/non-wildlife check, then using one of 3 specialist ones to determine the species. The collections page and achievement system which provides a richer experience for the user.

What we learned

How to prompt for specific features in the app, the basic functionalities of TensorFlow, how to use API calls that are rate limited, and focusing on the user experience. Offline functionalities are much greater with on device models.

What's next for Nature Radar

More features for the community tab, such as journal sharing and suggestions for what people should be on the lookout for during the week or specific areas.