Skip to content

README.md

Latest commit

 

History

History
259 lines (178 loc) · 7.71 KB

README.md

File metadata and controls

259 lines (178 loc) · 7.71 KB
Raw
Nano.AI Logo

🧠 Nano.AI

Your Complete Offline AI Companion for Android

Privacy-First • Powerful • Portable

Platform License Package

Run powerful AI models completely offline on your Android device. No internet required. No subscriptions. Complete privacy.

Download • Features • Documentation • Community

🌟 What is nano.ai?

nano.ai is a comprehensive AI ecosystem for Android that puts powerful AI capabilities directly in your pocket—completely offline. Chat with LLMs, generate images, use voice AI, and inject custom knowledge, all without an internet connection.

Why nano.ai?

  • 🔒 Privacy First - Your data never leaves your device in offline mode
  • 💪 Fully Featured - Chat, images, voice, and custom knowledge in one app
  • 🌐 Hybrid Mode - Switch between offline and cloud models seamlessly
  • 🆓 Free & Open - No subscriptions, no paywalls, Apache 2.0 licensed
  • 🔌 Extensible - Plugin system for unlimited customization

✨ Features

🤖 AI Chat (LLMs)

  • Offline Models: Run GGUF models (Llama, Mistral, Gemma, Phi) locally
  • Cloud Access: Connect to 100+ models via OpenRouter (GPT-4, Claude, Gemini)
  • Smart Streaming: Real-time token generation with context preservation
  • Conversation History: Full chat persistence with SQLite storage

🎨 Image Generation

  • Stable Diffusion 1.5: Generate images completely offline
  • Censored & Uncensored: Choose the model that fits your needs
  • Mobile Optimized: Runs on phones with 6GB+ RAM
  • Fast Generation: 30-90 seconds depending on device

🎙️ Voice AI

  • Text-to-Speech: 11 professional voices (American & British accents)
  • Speech-to-Text: Offline Whisper-powered recognition
  • Zero Latency: All processing happens on-device
  • Hands-Free: Perfect for driving or multitasking

🧠 Knowledge Injection (RAG)

  • Custom Data-Packs: Inject Wikipedia, docs, notes into AI context
  • No Retraining: Add knowledge without model retraining
  • Dynamic Mounting: Attach/detach knowledge bases on the fly
  • Universal Support: Works with both local and cloud models

🔌 Plugin System

  • Web Search: Real-time information retrieval
  • Web Scraper: Extract content from any URL
  • DataHub: Manage custom knowledge bases
  • Document Viewer: Analyze PDFs and text files
  • Extensible: Build your own plugins

📱 Screenshots

Chat Interface Model Selection Settings
Multi-modal AI conversations 100+ models available Complete customization

Screenshots coming soon

🚀 Quick Start Guide

Installation

Option 1: Direct Download

Download the latest APK from Releases and install on Android 8.0+ devices.

Option 2: Build from Source

# Clone repository
git clone https://github.com/NextGenXplorer/Nano.AI-Native.git
cd nano.ai

# Build with Gradle
./gradlew assembleDebug

# Install on connected device
./gradlew installDebug

Setup for Offline Use

  1. Load a Chat Model

    • Download a GGUF model from HuggingFace
    • Recommended: Llama-3-8B-Q4_K_M.gguf (4.5GB)
    • Import via Settings → Local Models

  2. Enable Image Generation

    • Download Stable Diffusion 1.5 model
    • Import via Settings → Image Models

  3. Activate Voice AI

    • TTS voices included by default
    • Download Whisper for STT via Settings → Voice Models

Setup for Cloud Use

  1. Get API key from OpenRouter.ai
  2. Enter key in Settings → API Configuration
  3. Access 100+ cloud models instantly

💻 System Requirements

Minimum (Cloud Models Only)

  • Android 8.0+ (API 26)
  • 4GB RAM
  • 2GB storage

Recommended (Local Models)

  • Android 10+
  • 6GB+ RAM (8GB preferred)
  • Snapdragon 8 Gen 1 or equivalent
  • 5GB+ storage

Optimal (Everything Offline)

  • Android 11+
  • 12GB+ RAM
  • Snapdragon 8 Gen 3 or equivalent
  • 10GB+ storage

🛠️ Technical Stack

Core Technologies:

  • Language: Kotlin + C++
  • UI: Jetpack Compose
  • Local Inference: llama.cpp (GGUF models)
  • Image Generation: Stable Diffusion C++
  • Voice: Sherpa-ONNX (TTS/STT)
  • Cloud API: Retrofit + OkHttp
  • Database: Room (SQLite)
  • Async: Kotlin Coroutines + Flow

Performance:

  • Quantized model support (Q4_K_M, Q5_K_S)
  • Context caching
  • Memory-mapped loading
  • NPU acceleration (where available)

🤝 Contributing

We welcome contributions! Here's how to get started:

  1. Fork the repository
  2. Create a feature branch (git checkout -b feature/amazing-feature)
  3. Commit your changes (git commit -m 'Add amazing feature')
  4. Push to branch (git push origin feature/amazing-feature)
  5. Open a Pull Request

Priority Areas

  • 🐛 Bug fixes and device compatibility
  • 📚 Documentation improvements
  • 🧪 Testing on various devices
  • 🔌 New plugin development
  • 🌍 Internationalization

📄 License

Distributed under the Apache 2.0 License. See LICENSE for details.

What you can do:

  • ✅ Commercial use
  • ✅ Modification
  • ✅ Distribution
  • ✅ Private use
  • ✅ Patent use

Requirements:

  • 📄 Include license and copyright notice
  • 📝 Document changes made

🙏 Acknowledgments

nano.ai is built on the shoulders of giants:

💬 Support & Community

❓ FAQ

Q: Will this drain my battery?
A: Local inference is power-intensive. For extended use, keep your device plugged in. Cloud mode uses minimal battery.

Q: How big are the model files?
A: GGUF models: 0.5GB-8GB. Stable Diffusion: ~2GB. Voice models: 50-500MB.

Q: Is my data really private?
A: In offline mode, absolutely nothing leaves your device. Verify in our open-source code.

Q: Can I use my own API keys?
A: Yes! Bring your own OpenRouter key. You control costs and usage.

Q: Does it support iOS?
A: Not currently. Android only due to platform constraints.

Built with ❤️ by NextGenXplorer

Privacy-first AI for everyone, everywhere

If nano.ai empowers your AI journey, please ⭐ star the repository!

Download • Report Bug • Request Feature

Powered by llama.cpp • Sherpa-ONNX • Stable Diffusion • OpenRouter • Jetpack Compose