LLM Launcher

Cross-platform CLI launcher for LLM agents (Aider) with support for Ollama (local) and OpenRouter (cloud) models.

Features

• 🚀 Cross-platform (Windows, macOS, Linux)
• 🤖 Support for Ollama local models
• ☁️ Support for OpenRouter cloud models
• 🎨 Interactive model selection with nice CLI interface
• ⚙️ Automatic dependency checking
• 🔐 Environment variable support via .env files
• 📦 TypeScript-based for type safety

Prerequisites

• Node.js >= 16
• Aider installed (pip install aider-chat)
• Ollama (optional, for local models)
• OpenRouter API key (optional, for cloud models)

Installation

Global Installation (Recommended)

npm install -g llm-launcher

Then run from anywhere:

llm
# or
llm-launcher

Local Development

git clone <repository-url>
cd llm-launcher
npm install
npm run build
npm link

Quick Start (without installing)

npx llm-launcher

Usage

Basic Usage

Simply run the command and follow the interactive prompts:

llm

The launcher will:

1. Check if Aider is installed
2. Prompt you to select a provider (Ollama or OpenRouter)
3. Show available models and let you select one
4. Launch Aider with your selected model

Using Ollama (Local Models)

1. Select "Ollama (Local models)" from the provider menu
2. The launcher will automatically start Ollama if it's not running
3. Select from your locally pulled models
4. Aider will launch with the selected model

First-time Ollama users:

# Pull a model first
ollama pull llama3.3
# or
ollama pull qwen2.5-coder
# or
ollama pull deepseek-coder

Using OpenRouter (Cloud Models)

1. Select "OpenRouter (Cloud API)" from the provider menu
2. Enter your OpenRouter API key (or set via environment variable)
3. Select from popular models or enter a custom model name
4. Aider will launch with the selected model

Get an OpenRouter API key: https://openrouter.ai

Configuration

Environment Variables

Create a .env file in your project directory or set environment variables:

# OpenRouter API key
OPENROUTER_API_KEY=your_api_key_here

# Ollama API base (optional, defaults to http://localhost:11434)
OLLAMA_API_BASE=http://localhost:11434

Supported Models

Ollama

Any model you've pulled locally via ollama pull <model-name>

OpenRouter

• Anthropic Claude (3.5 Sonnet, 3 Opus, 3 Sonnet, 3 Haiku)
• OpenAI GPT (GPT-4 Turbo, GPT-4, GPT-3.5 Turbo)
• Google Gemini (Pro 1.5, Pro)
• Meta Llama (3.1 70B, 3.1 8B)
• Mistral AI (Large, Medium)
• Qwen (2.5 72B)
• Custom models (enter any OpenRouter model identifier)

Development

Build

npm run build

Watch mode

npm run watch

Run in development

npm run dev

Publishing to npm

First-time Setup

1. Create an npm account at https://www.npmjs.com/signup
2. Login to npm from the command line:

npm login

Publishing a New Version

1. Update the version in package.json (or use npm version):

# Patch release (0.1.0 -> 0.1.1)
npm version patch

# Minor release (0.1.0 -> 0.2.0)
npm version minor

# Major release (0.1.0 -> 1.0.0)
npm version major

2. Publish to npm:

npm publish

The prepare script automatically builds the project before publishing.

Publishing a Beta/Pre-release

npm version prerelease --preid=beta
npm publish --tag beta

Users can install beta versions with:

npm install -g llm-launcher@beta

Architecture

src/
├── index.ts         # Main entry point and provider selection
├── types.ts         # TypeScript type definitions
├── utils.ts         # Utility functions (process checking, command execution)
├── ollama.ts        # Ollama provider implementation
└── openrouter.ts    # OpenRouter provider implementation

Why TypeScript?

• Type safety catches errors at compile time
• Better IDE support and autocomplete
• Self-documenting code with interfaces
• Easier refactoring and maintenance
• Still compiles to standard JavaScript

Troubleshooting

"Aider is not installed"

Install Aider:

pip install aider-chat

"Ollama is not installed"

Download and install Ollama from: https://ollama.ai

"No models found in Ollama"

Pull a model first:

ollama pull llama3.3

Windows-specific issues

Make sure you're running in PowerShell or Command Prompt with proper permissions.

License

MIT

Author

Thomas Powell

Contributing

Contributions welcome! Please feel free to submit issues or pull requests.