vincespeed/llm-runner-aio

LLM Runner AIO - All-In-One Local AI Platform

Portable Windows Desktop App — Run local AI models with a single click. No installation required.

Overview

LLM Runner AIO is a comprehensive, self-contained desktop application that bundles all the tools you need to run local AI models on your own hardware. No complex setup, no dependency hell — just download, run, and start chatting with AI locally.

What's Included

This package brings together four powerful open-source components into one seamless experience:

Component	Description	License
Open WebUI	Beautiful web interface for chatting with local LLMs (formerly Ollama WebUI)	MIT
llama.cpp	High-performance C++ inference engine for running LLMs locally	MIT
SearXNG	Privacy-respecting, metasearch engine for local web search	GPLv3
Vane	AI-powered browser automation and web interaction tool	Open Source

Features

• 🚀 Portable — No installer needed. Move the folder anywhere and it works.
• 🔒 100% Offline — All processing happens on your machine, no data leaves your computer
• 🎨 Modern UI — Clean, dark-themed interface with support for 8 languages
• ⚙️ Auto-Configuration — Automatic hardware detection and optimal settings
• 🔄 Multi-Service — Manages all AI services from one dashboard
• 💾 Persistent Database — Your chat history and settings are saved locally
• 📱 System Tray Integration — Runs quietly in the background
• 🌐 Port Configurable — Customize ports for all services
• 🖥️ Windows Autostart — Auto-start with Windows (registry + desktop shortcut)
• 🔧 Python venv Isolation — Dependencies installed in isolated virtual environment
• 🗄️ Manual Function Loading — Inject .json functions directly into OpenWebUI

How It Works

┌─────────────────────────────────────────────┐
│           LLM Runner AIO Launcher           │
├─────────────────────────────────────────────┤
│  ┌──────────┐  ┌──────────┐                │
│  │ SearXNG  │  │ llama.cpp│                │
│  │ :8080    │  │ :8000    │                │
│  └──────────┘  └──────────┘                │
│  ┌──────────┐  ┌──────────┐                │
│  │OpenWebUI │  │  Vane    │                │
│  │ :3000    │  │ :3001    │                │
│  └──────────┘  └──────────┘                │
└─────────────────────────────────────────────┘

1. llama.cpp runs the AI model inference engine
2. Open WebUI provides the chat interface at http://localhost:3000
3. SearXNG enables local web search at http://localhost:8080
4. Vane handles browser automation at http://localhost:3001

System Requirements

Requirement	Minimum	Recommended
OS	Windows 10/11	Windows 11
RAM	8 GB	16 GB+
VRAM	N/A	4 GB+
Python	3.11+ (system)	3.11+ (system)
Node.js	18+ (system)	20+ (system)

Note: Python 3.11 and Node.js must be installed on your system before running the app.

Quick Start

First Time Setup

1. Download or clone the repository:

   git clone https://github.com/nafiozdemir231-prog/llm-runner-aio.git
   cd llm-runner-aio
   

2. Install system dependencies:

   • Python 3.11+ (from https://python.org)
   • Node.js 18+ (from https://nodejs.org)

3. Run the launcher:

   # Windows — creates venv and starts the app
   .\run.bat
   
   # Or for silent background mode:
   cscript //Nologo electron\launch_app.vbs
   

4. Open your browser and go to http://localhost:3000

5. Start chatting with your local AI!

Development Mode

# Install Electron dependencies
npm install

# Start with DevTools auto-opened
npm start -- --dev

Windows Autostart

The app automatically configures itself on every launch:

• Registry entry → HKCU\...\Run\LLMRunnerAIO points to electron/launch_app.vbs
• Desktop shortcut → Created via PowerShell
• Startup folder → .lnk shortcut created in %APPDATA%\Microsoft\Windows\Start Menu\Programs\Startup

All paths are resolved dynamically — move the folder and autostart still works.

Building Distribution Package

npm run build:win    # Windows x64 NSIS installer

Built packages will be in the dist/ directory.

Supported Languages

The interface supports 8 languages:

• 🇹🇷 Turkish (Türkçe)
• 🇬🇧 English
• 🇪🇸 Spanish (Español)
• 🇩🇪 German (Deutsch)
• 🇫🇷 French (Français)
• 🇵🇹 Portuguese (Português)
• 🇨🇳 Chinese (中文)
• 🇯🇵 Japanese (日本語)

Development Guide

Architecture Overview

LLM-Runner-AIO/
├── electron/
│   ├── main.js              # Main process (lifecycle, IPC, autostart, cleanup)
│   ├── preload.js           # Context bridge (secure API exposure)
│   ├── launch_app.vbs       # Silent launcher for Windows autostart
│   └── create_shortcut.py   # Desktop shortcut creator
├── src/
│   ├── index.html           # UI structure
│   ├── css/style.css        # Modern transparent/minimal styling
│   ├── renderer.js          # Renderer process (DOM, tabs, logs)
│   ├── lang/                # 8 language files (en, tr, es, de, fr, pt, zh, ja)
│   └── assets/              # Icons, images
├── venv/                    # Python virtual environment (auto-created)
├── node_modules/            # Node.js + Electron dependencies
├── launcher/                # PyQt6 reference (legacy)
├── searxng/                 # SearXNG configuration & data
├── openwebui/               # OpenWebUI runtime directory
├── picoding/                # MCP Advisor settings
├── run.bat                  # Interactive launcher (creates venv, installs deps)
├── run_silent.bat           # Silent mode for autostart
├── Vane/                    # Vane Next.js app
├── gpu*.ini                 # Hardware-specific model presets
├── model_urls.json          # Model download URLs
└── requirements.txt         # Python dependencies

Key Technologies

Component	Technology
Desktop Framework	Electron v28+
UI	HTML5, CSS3, Vanilla JS
Process Management	child_process.spawn() + tree-kill
Database	SQLite (better-sqlite3)
Python Isolation	venv (auto-created by run.bat)
Autostart	Registry + VBS launcher + .lnk shortcuts

Adding New Features

1. New Tab: Add to src/index.html, create in src/tabs/, register in renderer.js
2. New IPC Channel: Add to preload.js (expose) and main.js (handler)
3. New Language: Add to src/lang/*.json (follow existing format)
4. New Server: Update startServer() in main.js spawn logic

Data Privacy & Security

• ✅ No Cloud Dependencies — Everything runs locally
• ✅ No Telemetry — No data is sent anywhere
• ✅ Local Database — Chat history stored only on your machine
• ✅ No Account Required — No registration or login needed
• ✅ Portable — No registry entries left behind when moved

License

This package is distributed under the MIT License. See LICENSE.txt for full details.

Individual components retain their original licenses:

• llama.cpp: MIT
• Open WebUI: MIT
• SearXNG: GPLv3

Credits

This project would not be possible without the incredible work of:

• Georgi Gerganov — llama.cpp
• Open WebUI Team — Open WebUI
• SearXNG Contributors — SearXNG
• All open-source contributors who make local AI accessible

Support

For questions, bug reports, or feature requests:

• Website: https://aihublocal.com/llm-runner-aio
• Ko-fi: Support the project at https://ko-fi.com/vincespeed

Disclaimer

This software is provided "as is" without warranty of any kind. The developer (AIHubLocal) is not responsible for any data loss, system issues, or hardware strain resulting from use. Running local LLMs consumes significant system resources — users assume all risks associated with usage.