Kolosal CLI

AI-powered command-line workflow tool for developers

Installation • Quick Start • Features • Documentation • Contributing

Kolosal CLI is a powerful command-line AI workflow tool that enhances your development workflow with advanced code understanding, automated tasks, and intelligent assistance. Based on Qwen Code models, it provides seamless integration with your existing development environment.

Features

• Code Understanding & Editing - Query and edit large codebases beyond traditional context window limits
• Workflow Automation - Automate operational tasks like handling pull requests and complex rebases
• Enhanced Parser - Optimized specifically for code-oriented models
• Vision Model Support - Automatic detection and multimodal analysis of images in your input
• Flexible Model Support - Works with OpenAI-compatible APIs and Hugging Face models

Installation

Quick Install (Recommended)

Install KolosalCode with a single command:

# macOS or Linux
curl -fsSL https://raw.githubusercontent.com/KolosalAI/kolosal-cli/main/install.sh | bash

# Linux (with sudo)
curl -fsSL https://raw.githubusercontent.com/KolosalAI/kolosal-cli/main/install.sh | sudo bash

The installer automatically detects your operating system and installs the appropriate package.

Alternative Installation Methods

Install from npm

Requires Node.js version 20 or higher.

npm install -g @kolosal-ai/kolosal-ai@latest
kolosal --version

Manual Download

macOS:

# Download and install the .pkg
curl -LO https://github.com/KolosalAI/kolosal-cli/releases/download/v0.1.0-pre/KolosalCode-macos-signed.pkg
sudo installer -pkg KolosalCode-macos-signed.pkg -target /

Linux (Debian/Ubuntu):

# Download and install the .deb
wget https://github.com/KolosalAI/kolosal-cli/releases/download/v0.1.0-pre/kolosal-code_0.0.14_amd64.deb
sudo dpkg -i kolosal-code_0.0.14_amd64.deb
sudo apt-get install -f  # Fix dependencies if needed

Build from source

git clone https://github.com/KolosalAI/kolosal-cli.git
cd kolosal-cli
npm install
npm run build
npm link  # Optional: install globally

See CONTRIBUTING.md for detailed build instructions.

Verify Installation

kolosal --version
kolosal --help

For detailed installation instructions and troubleshooting, see docs/INSTALLATION.md.

Homebrew (Coming Soon)

brew install kolosal-ai

Quick Start

# Start Kolosal CLI
kolosal

# Example commands
> Explain this codebase structure
> Help me refactor this function
> Generate unit tests for this module

Authentication Options

1. OpenAI Compatible API

Set environment variables or create a .env file:

export OPENAI_API_KEY="your_api_key_here"
export OPENAI_BASE_URL="your_api_endpoint" 
export OPENAI_MODEL="your_model_choice"

2. Hugging Face LLM Models

Configure Hugging Face models for local or API-based inference:

export HF_TOKEN="your_huggingface_token"
export HF_MODEL="your_model_name"

Note: Kolosal CLI may issue multiple API calls per cycle, which can result in higher token usage. We're actively optimizing API efficiency.

Configuration

Session Management

Configure token limits in .kolosal/settings.json:

{
  "sessionTokenLimit": 32000
}

Vision Models

Kolosal CLI automatically detects images and switches to vision-capable models. Configure the behavior:

{
  "experimental": {
    "vlmSwitchMode": "once",  // "once", "session", "persist"
    "visionModelPreview": true
  }
}

Usage Examples

Code Analysis & Understanding

> Explain this codebase architecture
> What are the key dependencies and how do they interact?
> Find all API endpoints and their authentication methods
> Generate a dependency graph for this module

Code Development & Refactoring

> Refactor this function to improve readability and performance
> Convert this class to use dependency injection
> Generate unit tests for the authentication module
> Add error handling to all database operations

Workflow Automation

> Analyze git commits from the last 7 days, grouped by feature
> Create a changelog from recent commits
> Find and remove all console.log statements
> Check for potential SQL injection vulnerabilities

Commands & Shortcuts

Session Commands

• /help - Display available commands
• /clear - Clear conversation history
• /compress - Compress history to save tokens
• /stats - Show current session information
• /exit or /quit - Exit Kolosal CLI

Keyboard Shortcuts

• Ctrl+C - Cancel current operation
• Ctrl+D - Exit (on empty line)
• Up/Down - Navigate command history

Development & Contributing

See CONTRIBUTING.md to learn how to contribute to the project.

Distribution

Windows Package

Build a single executable and installer:

npm run bundle               # builds bundle/gemini.js
npm run build:win:exe        # creates dist/win/kolosal.exe
iscc installer/kolosal.iss   # creates installer

macOS Package

Build a standalone .pkg installer with bundled Node.js:

npm run build:mac:pkg

Creates dist/mac/KolosalCode-macos-signed.pkg - a self-contained package that requires no Node.js installation.

Linux Packages

Build Linux packages (.deb, .rpm, and .tar.gz):

# Build all formats
npm run build:linux

# Build specific format
npm run build:linux:deb      # Debian/Ubuntu package
npm run build:linux:rpm      # Red Hat/Fedora package  
npm run build:linux:tar      # Universal tarball

# Or use the convenience script
bash scripts/clean-build-linux.sh [all|deb|rpm|tar]

Creates packages in dist/linux/ with bundled Node.js - no separate runtime installation needed.

See docs/LINUX-PACKAGING.md for detailed Linux packaging documentation.

Acknowledgments

Kolosal CLI is built upon and adapted from the Qwen Code project, which is licensed under the Apache License, Version 2.0. We acknowledge and appreciate the excellent work of the Qwen development team.

This project also incorporates concepts and approaches from Google Gemini CLI. We appreciate the foundational work that made this project possible.

License

Licensed under the Apache License, Version 2.0. See LICENSE for the full license text.

This product includes software developed from the Qwen Code project, which is also licensed under the Apache License, Version 2.0.