jegly

jegly / OfflineLLM

Public

A privacy-first Android chat app that runs large language models entirely on-device. No internet, no cloud, no tracking. Built with Kotlin, Jetpack Compose, and llama.cpp with optimized ARM NEON/SVE inference.

jegly.xyz ai-android-app android android-ai-app android-llm edge-ai
10
1
100% credibility
Found Apr 06, 2026 at 10 stars -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
Kotlin
AI Summary

An open-source Android app for running AI language models completely offline to enable private, internet-free chat conversations.

How It Works

1
📱 Discover OfflineLLM

You find this private AI chat app on GitHub and download the ready-to-use file for your Android phone.

2
🔧 Install the app

Allow installs from your file manager, tap the file, and add it to your phone like any other app.

3
🧠 Set up your AI companion

Open the app, follow the welcome guide to grab a smart brain file from a model site or use the one included, and pick an assistant style like tutor or coder.

4
💬 Start your first chat

Type a message in the clean chat screen, watch responses stream in real-time, and build conversations that stay just between you and your phone.

5
⚙️ Make it yours

Tweak colors, voices, creativity sliders, or lock it with your fingerprint to match your style and keep it secure.

🔒 Chat privately forever

Relax knowing every word stays offline on your device—no tracking, no cloud, pure privacy in your pocket.

Sign up to see the full architecture

4 more

Sign Up Free

Star Growth

See how this repo grew from 10 to 10 stars Sign Up Free
Repurpose This Repo

Repurpose is a Pro feature

Generate ready-to-use prompts for X threads, LinkedIn posts, blog posts, YouTube scripts, and more -- with full repo context baked in.

Unlock Repurpose
AI-Generated Review

What is OfflineLLM?

OfflineLLM is a privacy-first Android chat app that runs large language models entirely on-device using Kotlin, Jetpack Compose, and llama.cpp with ARM-optimized inference. It delivers offline LLM chats without internet, cloud, or tracking—your conversations stay local on your privacy-first Android phone. Import any GGUF model, get streaming responses at 25-60 tokens/second on midrange devices, and manage multiple auto-titled threads.

Why is it gaining traction?

This offline LLM GitHub project stands out with zero network permissions, full sampling controls (temperature, top-p/k, repeat penalty), and features like TTS readout, biometric lock, and encrypted backups. Developers dig the seamless model import from Hugging Face, system prompts for coding/tutoring, and AMOLED-black theming built for Android. No cloud dependency means instant privacy without sacrificing speed on ARM hardware.

Who should use this?

Android enthusiasts seeking a privacy-first browser for Android alternative for AI chats, or devs prototyping local inference on phones with 4GB+ RAM. Ideal for users tired of cloud LLMs leaking data, like journalists needing secure offline note-taking or travelers wanting chat without roaming charges. Tinkerers evaluating GGUF models on real hardware will appreciate the benchmark tools and exportable JSON histories.

Verdict

Grab the APK from releases if you need on-device AI now—it's polished for an early project with solid docs and performance. With just 10 stars and 1.0% credibility score, treat it as alpha: fork and contribute to push maturity, but it already beats bloated cloud apps for privacy-first Android use.

(198 words)

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.