Skip to main content

​
Introduction

The Cactus CLI provides a complete toolkit for downloading, converting, running, and testing AI models on-device. All commands are accessible through the cactus executable after installation.

​
Installation

​
From Source

git clone https://github.com/cactus-compute/cactus && cd cactus
source ./setup

​
Prerequisites (Linux)

On Ubuntu/Debian systems, install required dependencies: 
sudo apt-get install python3 python3-venv python3-pip cmake \
  build-essential libcurl4-openssl-dev

​
Available Commands

cactus auth

Manage Cloud API key for fallback

cactus run

Open interactive playground with any model

cactus transcribe

Live microphone or file transcription

cactus download

Download and cache models locally

cactus convert

Convert models with LoRA merge support

cactus build

Build native libraries for iOS, Android, Flutter

cactus test

Run unit tests and benchmarks

cactus clean

Remove all build artifacts

​
Quick Start

# Run a model in interactive mode
cactus run qwen-2.5-1.5b

# Transcribe audio from your microphone
cactus transcribe

# Download a model for offline use
cactus download llama-3.2-1b

# Build for iOS
cactus build --apple

​
Getting Help

View all available commands and flags: 
cactus --help

​
Common Flags

Many commands share common flags:
  • --precision INT4|INT8|FP16 - Model quantization level (default: INT4)
  • --token <token> - HuggingFace API token for gated models
  • --reconvert - Force reconversion from source weights

​
Next Steps

Run Models

Learn how to run models interactively

Model Library

Browse available models