I’m building a personal assistant app that lives entirely on-device and responds to spoken commands. The core of the project is a voice interface capable of: • sending SMS or WhatsApp-style text messages and initiating phone calls • creating reminders and alarms through natural language • toggling or adjusting device settings such as Wi-Fi, Bluetooth, brightness and volume Everything must run standalone inside the app—no hooks into external cloud services or third-party platforms. Using the standard Android APIs for telephony, alarms, and settings is fine; the assistant just shouldn’t rely on outside apps to complete an action. I’m happy with any modern Android stack (Kotlin, Java, Jetpack Compose, ML Kit, TensorFlow Lite for on-device speech), as long as the voice recognition is fast, accurate, and operates locally. Deliverables • Source code with clear comments • A runnable APK for testing on Android 10+ • Brief setup notes explaining how to train/add new voice intents Once the basics work smoothly, I may extend the project, so well-structured, maintainable code is essential.