Project Name
Zebra Alex
Your Name or Team
Nick Cadenhead
Repository/Code Link
https://github.com/nicholasc69/InvestecBankingWidget
Live Demo (if any)
No response
Description
An offline-first Android application and Home Screen Widget for Investec Private Banking. It integrates secure local database caching, biometric security, a Jetpack Glance home screen widget, and an on-device AI assistant ("Alex") powered by Google LiteRT (formerly TensorFlow Lite) using Gemma.
Screenshots or GIFs
How to Run/Use It
⚙️ Getting Started
1. Prerequisites
- Android Studio (Koala or newer recommended).
- Android SDK 35/36 target compatibility.
- An emulator or physical device supporting Biometric Authentication.
2. Set Up Local Gemma Model (LiteRT)
The AI assistant runs a local language model. You must supply a compatible LiteRT model file:
- Obtain the
gemma-4-E2B-it.litertlm model. Hugging Face
- Push the model to the target device's local tmp directory via ADB:
adb push path/to/gemma-4-E2B-it.litertlm /data/local/tmp/
3. Open API Credentials Setup
The application comes preconfigured to run in Sandbox Mode using open-access sandboxed credentials.
To connect to your live Investec accounts:
- Open the app and tap the Settings (gear) icon in the top right.
- Disable Developer Sandbox Mode.
- Supply your Client ID, Client Secret, and x-api-key obtained from the Investec Developer Portal.
- Tap Apply & Sync to fetch your actual accounts and transactions securely.
🧪 Testing
The project is fully integrated with:
- Compose UI Tests and Robolectric for local behavior verification.
- Roborazzi for screenshot testing.
- To run tests:
Code of Conduct
Project Name
Zebra Alex
Your Name or Team
Nick Cadenhead
Repository/Code Link
https://github.com/nicholasc69/InvestecBankingWidget
Live Demo (if any)
No response
Description
An offline-first Android application and Home Screen Widget for Investec Private Banking. It integrates secure local database caching, biometric security, a Jetpack Glance home screen widget, and an on-device AI assistant ("Alex") powered by Google LiteRT (formerly TensorFlow Lite) using Gemma.
Screenshots or GIFs
How to Run/Use It
⚙️ Getting Started
1. Prerequisites
2. Set Up Local Gemma Model (LiteRT)
The AI assistant runs a local language model. You must supply a compatible LiteRT model file:
gemma-4-E2B-it.litertlmmodel. Hugging Face3. Open API Credentials Setup
The application comes preconfigured to run in Sandbox Mode using open-access sandboxed credentials.
To connect to your live Investec accounts:
🧪 Testing
The project is fully integrated with:
./gradlew testCode of Conduct