Arandu

Arandu is a Windows desktop application for managing llama.cpp models and servers without manually handling DLL files or command-line arguments.

Version: 0.5.5-beta
Platform: Windows only

Features

• Model Management - Organize and launch .gguf models with custom arguments
• HuggingFace Integration - Search and download models directly from HuggingFace
• AI Model Tracker - Browse trending AI models from HuggingFace with filtering and stats
• Llama.cpp Backend Management - Auto-download llama-server backends from GitHub releases
• Flexible Launch Options - Internal (integrated UI) or External (native llama.cpp UI) modes
• Argument Presets - Save and reuse common launch configurations per model
• Hardware Monitoring - Real-time RAM and VRAM usage indicators
• Desktop Organization - Sort models by name, size, architecture, or quantization
• Color Themes - Multiple UI themes with customizable backgrounds

Quick Start

Prerequisites

• Windows 10/11
• Rust installed
• Visual Studio Build Tools (for Windows)

Development

cd backend
cargo tauri dev

Building

cd backend
cargo tauri build
# Output: backend/target/release/Arandu.exe

Documentation

See AGENTS.md for detailed developer documentation including architecture, API reference, and common issues.

For AI Agents/Developers

⚠️ MANDATORY: New agents MUST follow the onboarding workflow before making any changes.

Critical Documentation (read in order):

1. WORKING_DIRECTORY_WARNING.md - CRITICAL working directory info
2. AGENTS.md - Architecture, patterns, and how-to guides
3. THIS-PROJECTS-CURRENT-STATE.md - Current status, recent bugs, build info

Knowledge Base Memory System: This project uses the nowledge-mem memory system to store:

• File locations (search: "Arandu Complete File Location Reference")
• Common development patterns (search: "Arandu Common Development Patterns")
• Critical gotchas and issues (search: "Arandu Critical Gotchas")
• Feature implementations (search: "Arandu [feature-name]")

Agent Onboarding:

• Check memory FIRST before using shell commands to find files
• Search for "MANDATORY: Arandu Agent Onboarding Workflow" for complete workflow
• Always verify working directory is H:\Ardanu Fix\Arandu-maxi\ (NOT C drive)
• Build and test before claiming work is complete

Download

Pre-built releases: https://github.com/fredconex/Arandu/releases

Screenshots

Roadmap

• Code cleanup and organization
• New features (coming soon)

Support Development