⚡ LocalInter — Interface for AI Model Chat on Android (Root Required)

LocalInter = Local Interface

This project is the UI / front-end layer for the local AI model setup

Before You Use This app, If you haven’t set up your offline AI model on your device yet, STOP here and follow the installation guide:

👉 Cross Compile Bins & Exe's (GGUF + llama.cpp on Android) https://github.com/1ntern0t/android-offline-ai

That repo shows you how to: cross-compile llama.cpp for your device push the GGUF model to /data/local/tmp/llama/ enable root execution of llama-cli test the model fully offline (no Wi-Fi/data)

Once that setup is complete and the model runs successfully, this app makes everything easier.

What This App Does: Run your AI model (via llama-cli) Run Python code directly (via Chaquopy) Toggle between AI mode and Python mode Chat with the model without opening a shell Avoid typing root commands manually Automatically pipe inputs to root llama-cli behind the scenes Everything is handled in the background:

JNI bridges Root shell commands LD_LIBRARY_PATH Model path + prompt formatting Python interpreter execution

You just type your prompt → the app handles the rest.

Why This Exists

Using su, terminal commands, and manual flags is great for testing — but annoying for daily interaction.

This app: removes all terminal complexity replaces commands with a normal text field shows clean output in a UI allows fast toggling between Python + LLaMA It’s the front-end brain for your local AI Models

Requirements Rooted Android device Binaries and Executables must be Cross Compiled (from the guide above) GGUF file placed in /data/local/tmp/llama/ llama-cli compiled using your device's architecture Android device running minSdk 26+