Fieldnote

Build a mobile app called Fieldnote: log birds by voice while your hands stay on the binoculars.

What it does

Say “two great tits, one robin, by the lake” and Fieldnote logs it with time and location. On-device speech recognition, works without signal, exports to eBird.

Stack

Use Swift, Core ML. The original was built with Cursor and Claude in 4 weekends, for about $25 in AI usage.

How it fits together

• A mobile app with three or four screens and a tab bar
• A backend for sync and accounts
• Push notifications only where they help

Build it in this order

1. Design the main screen for the youngest or least technical user
2. Build the screens with sample data
3. Add accounts and sync
4. Add the one notification that matters
5. Test it with a real person and fix what confuses them

A prompt that worked: Match messy speech to species

Write a Swift function that maps a transcribed phrase to species from a local list of 600 European birds. Handle plurals, counts spoken as words, common nicknames and mishearings (e.g. ‘great tit’ heard as ‘grey tit’). Return the top match with a confidence, and the second best if it's close.

Watch out for

• Big tap targets and short labels; people use it with one hand
• It should work offline and sync later

Done when

• The main task takes three taps or fewer
• It works offline
• A first-time user gets it without help

Before writing code, ask me up to five questions about who this outdoors mobile app is for and what the first version must do. Then work in small steps: after each one, tell me how to try it and wait for my feedback.