Show HN: Kibo, AI expense tracker for Android with AI categorization

Hi HN,

I’ve been working nights and weekends on my first Android app, and last week I finally launched it. It’s called Kibo, a simple, privacy-focused expense tracker that uses AI and OCR to make logging expenses less painful.

The problem I wanted to solve: I tried a bunch of budgeting apps, but most were either slow, bloated, full of ads, or locked basic features behind subscriptions. I wanted something I’d actually use every day — fast, lightweight, and private.

What Kibo does: • AI suggests categories + emojis instantly from the expense name • OCR reads receipts and fills in transaction details automatically • Clean Material You design with light/dark/system themes • Free tier includes 20 transactions/month with unlimited custom categories • Secure cloud sync (end-to-end encrypted)

Tech stack: Built with Kotlin + Jetpack Compose, Firebase for auth/storage, and MLKit for OCR. AI categorization uses a lightweight NLP model fine-tuned on category datasets.

I’d love feedback on both the UX and the technical side. How would you approach improving discoverability of a niche consumer app like this? Also open to suggestions on the architecture — right now everything runs client-side except sync.

Play Store link: https://play.google.com/store/apps/details?id=com.kibo.app

Thanks for checking it out — happy to dive into the code or design decisions if anyone’s curious.