Skip to content

SamirSengupta/PrometheusAI-Android

Repository files navigation

PrometheusAI-Android

Welcome to the official repository for PrometheusAI-Android. This project allows you to run a powerful AI model locally on your Android device using ONNX Runtime.

🚀 Features

  • Local Inference: Runs entirely on-device without internet access.
  • Privacy Focused: No data leaves your device.
  • Customizable: Built on open-source technologies.

🛠️ Prerequisites

Before you begin, ensure you have the following installed:

  • Android Studio (Latest version recommended)
  • Git installed on your machine.
  • A decent Android device (or emulator) with at least 4GB - 8GB of RAM for smooth performance.

📥 Installation & Setup

Follow these steps to set up the project locally:

1. Clone the Repository

Open your terminal or command prompt and run:

git clone https://github.com/SamirSengupta/PrometheusAI-Android.git
cd PrometheusAI-Android

2. Download Model Files (CRITICAL STEP)

The AI model files (model.onnx and tokenizer.json) are too large to be included in this repository. You must download them manually and place them in the correct folder for the app to work.

  1. Download the files from the link below: 👉 [INSERT YOUR MODEL DOWNLOAD LINK HERE] (e.g., Google Drive link containing model.onnx and tokenizer.json)

  2. Locate the assets folder in the project directory: app/src/main/assets/

  3. Move the downloaded files into that folder.

    • Ensure the file names are exactly model.onnx and tokenizer.json.
    • The final path should look like:
      • .../PrometheusAI-Android/app/src/main/assets/model.onnx
      • .../PrometheusAI-Android/app/src/main/assets/tokenizer.json

3. Open in Android Studio

  1. Launch Android Studio.
  2. Select File > Open.
  3. Navigate to the cloned PrometheusAI-Android folder and select it.
  4. Wait for Gradle to sync (this may take a few minutes).

🏗️ How to Build Your Own PrometheusAI

Once the project is set up, you can build and run it on your device.

Running on a Device

  1. Connect your Android device via USB.
  2. Enable Developer Options and USB Debugging on your phone.
  3. In Android Studio, select your device from the dropdown menu in the toolbar.
  4. Click the green Run button (▶️) or press Shift + F10.

Building an APK

If you want to share the app or install it manually:

  1. Go to Build > Build Bundle(s) / APK(s) > Build APK(s).
  2. Once finished, a notification will appear. Click locate to find the app-debug.apk file.
  3. Transfer this file to your phone and install it.

🔧 Customization

Want to make it your own?

  • Change the Model: You can replace model.onnx with other ONNX-compatible LLMs (ensure you update the code to handle different architectures if needed).
  • Modify UI: Edit the files in app/src/main/res/layout to change the look and feel.
  • Logic: The main logic resides in MainActivity.kt.

❓ Troubleshooting

  • Build Failed (OOM Error): If the build fails due to memory, try increasing the heap size in gradle.properties.
  • App Crashes on Start: Ensure you have successfully placed model.onnx and tokenizer.json in the assets folder. The app cannot run without them.

Built with ❤️ by Samir Sengupta

About

Offline LLM which runs locally on your android device

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages