xxrickyxx/AiLo_UnlimitedLLama

🦙 Unlimited Llama - AI Desktop Assistant AiLo Core

A complete AI desktop assistant with chat, web search, speech synthesis, and OCR.

---

✨ Features

• 💬 Smart chat with local GGUF models
• 🌐 Integrated web search for up-to-date information
• 🔊 Text-to-Speech (TTS) and Speech Recognition (STT)
• 📷 OCR to extract text from images
• 💾 Advanced session management
• 🎛️ Supports any LLM model size
• 🔌 OpenAI-compatible API server
• 📤 Export in JSON, TXT, and Markdown
• 🌐 Integrated distribuited computing

---

📦 Download

---

🚀 Quick Start Guide

First Launch

1. Load a model → 🤖 Model → 📁 Load Model
2. Start chatting → type in the box below and press Enter
3. Sessions are saved automatically

---

🔍 Web Search

• Enable/disable using the 🌐 Web Search toggle
• Automatically searches for news, recent info, or local data
• Displays the sources used

🔊 Speech Synthesis (TTS)

• Enable via 🔊 TTS in the sidebar
• The assistant reads responses aloud
• Use 🔇 STOP to interrupt

🎤 Speech Recognition

• 🎤 Voice Input for single input
• 🎤 Start Listening for continuous mode

📷 OCR from Images

• Click 📷 Image OCR
• Select an image (PNG, JPG, etc.)
• Extracted text is automatically inserted into the chat

---

🛠️ Troubleshooting

❌ “Model not found”

• Make sure the GGUF file is in the /models folder
• Verify the file format is .gguf
• Check that you have enough disk space

❌ “Tesseract not found”

• Install Tesseract OCR following the instructions below
• Restart the application after installation

---

---

⚙️ Configuration Memory Optimization Memory Mapping (MMAP) What it does: Maps model directly from disk instead of loading entirely into RAM

Benefits: Reduces RAM usage by up to 70%, faster startup

Use when: Limited RAM, large models (>7GB)

Performance: Slightly slower inference, much less RAM usage

Memory Locking (MLOCK) What it does: Locks model in RAM preventing swap to disk

Benefits: Maximum performance, consistent response times

Use when: Abundant RAM, performance-critical applications

Performance: Fastest inference, permanent RAM occupation

⚙️ System Requirements

Minimum

• OS: Windows 10/11, macOS 10.15+, Linux (Ubuntu 18.04+)
• RAM: 8 GB (16 GB recommended)
• Disk Space: 2 GB + space for models
• CPU: Modern 64-bit processor

Recommended

• RAM: 16 GB+ for large models
• GPU: NVIDIA/AMD with CUDA or Metal (optional)
• Disk Space: 10 GB+ for large models

---

🔧 Installation

1. Install Tesseract OCR (Required for OCR)

Windows

# Using Chocolatey (recommended)
choco install tesseract