β¨ Run Offline AI Chatbots on Android with ChatNONET β No Internet Needed!
π§ ChatNONET is a fully offline Android app that uses quantized LLMs to answer your questions β fast, private, and without internet access.
π What is ChatNONET?
ChatNONET is my open-source Android app that brings the power of local large language models (LLMs) to your device β no internet required. It's powered by my own fine-tuned models called NONET, built for fast, accurate, direct question answering.
Whether you're offline, privacy-conscious, or just love tinkering with AI, this app is for you.
π§ How Does It Work?
ChatNONET uses:
- π Fine-tuned LLMs (based on Smollm & LLaMA 3.2)
- βοΈ Quantized to
.gguf
format using llama.cpp - π± Integrated into a native Android app (built in Android Studio)
- π§ Supports multiple model sizes from 135M to 3B parameters
π¦ Download and Try
You can also run the model in Python using llama.cpp
:
./llama-cli -m ./ChatNONET-300m-tuned-q8_0.gguf -p "You are ChatNONET AI assistant." -cnv
π Model Variants
Model Name | Base | Size |
---|---|---|
ChatNONET-135M | Smollm | 135M |
ChatNONET-300M | Smollm | 300M |
ChatNONET-1B | LLaMA 3.2 | 1B |
ChatNONET-3B | LLaMA 3.2 | 3B |
All models are quantized using q8_0
for smooth performance.
π‘ Why Offline?
- π Privacy β No data leaves your phone.
- π Connectivity β Works even in remote areas.
- π Speed β No server, no lag.
- π± Control β You choose the model and usage.
π» Open Source on GitHub
Want to build your own offline AI app?
π Check out the full ChatNONET Android source code on GitHub:
https://github.com/Mca-Tech/ChatNONET
πββοΈ Who Is This For?
- Android developers interested in AI
- Hobbyists & privacy advocates
- Students learning about local LLMs
- Anyone who wants ChatGPT-like power without the cloud
I built ChatNONET to explore how far local AI can go β and the results are exciting! You donβt need powerful servers or always-on internet. You just need your curiosity, an Android phone, and a bit of model magic.
π Like this project?
- Star the repo β
- Share your thoughts π¬
- Try the app!
Thanks for reading!
β Michael Cobol Agan (McaTech)
Top comments (0)