QLLM: Quantalogic Large Language Model CLI & AI Toolbox 🚀

Table of Contents

...

Table of Contents

1. Introduction
2. Features
3. Installation
4. Configuration
5. Usage
6. Advanced Features
7. Command Reference
8. Examples
9. Troubleshooting
10. Contributing
11. License
12. Acknowledgements

1. Introduction

Welcome to QLLM CLI, the ultimate command-line interface for seamless interaction with Large Language Models (LLMs). Crafted with passion by @quantalogic, QLLM CLI revolutionizes the way you engage with AI, offering a unified platform that supports multiple providers and empowers users with unparalleled flexibility.

In today's AI-driven landscape, QLLM CLI emerges as a game-changing solution for developers, researchers, and AI enthusiasts alike. Whether you're integrating AI into your workflow, exploring the frontiers of language models, or simply engaging in thought-provoking conversations with AI, QLLM CLI provides the robust toolkit you need.

Key Highlights:

• Multi-provider support (OpenAI, Anthropic, and more)
• Rich, interactive chat experiences with advanced conversation management
• Efficient one-time question answering for quick insights
• Cutting-edge image input capabilities for visual analysis tasks
• Fine-grained control over model parameters for tailored responses
• Comprehensive configuration options for a personalized experience

Embrace the future of AI interaction with QLLM CLI – your gateway to boundless possibilities in the world of language models.

2. Features

QLLM CLI boasts an impressive array of features designed to elevate your AI interaction experience:

1. 🌐 Multi-provider Support: Effortlessly switch between LLM providers like OpenAI and Anthropic, leveraging the unique strengths of various models and facilitating comparative analysis.

2. 💬 Interactive Chat Sessions: Immerse yourself in dynamic, context-aware conversations with LLMs, complete with robust conversation history management.

3. ❓ One-time Question Answering: Quickly obtain answers to standalone queries without the need for a full chat session.

4. 🖼️ Image Input Support: Analyze images from multiple sources:

   • Local files on your system
   • URLs pointing to online images
   • Images from your clipboard
   • Screenshots captured directly through the CLI

5. 🎛️ Customizable Model Parameters: Fine-tune AI behavior with adjustable settings:

   • Temperature
   • Max tokens
   • Top P
   • Frequency penalty
   • Presence penalty
   • Stop sequences

6. 🗂️ Conversation Management: Efficiently save, list, load, and delete chat histories for easy reference and continuation of previous interactions.

7. 📋 Provider and Model Listing: Easily view available providers and their associated models to make informed choices.

8. 🔄 Streaming Responses: Experience real-time output for long-form content, enhancing interactivity.

9. 💾 Output to File: Save AI responses directly to your filesystem for future reference or processing.

10. ⚙️ Configurable Settings: Tailor your QLLM CLI experience with a robust configuration system for managing defaults and API keys.

11. 🖥️ Cross-platform Compatibility: Enjoy a consistent experience across Windows, macOS, and Linux.

12. 📸 Screenshot Capture: Take screenshots directly from the CLI for immediate AI analysis.

13. 🎨 Syntax Highlighting: Benefit from colorized output for improved readability and a more pleasant user experience.

14. 🛡️ Error Handling and Validation: Robust error checking and input validation ensure smooth operation and helpful error messages.

15. 🧩 Extensible Architecture: Modular design facilitates easy addition of new features and providers in the future.

3. Installation

To embark on your QLLM CLI journey, ensure you have Node.js (version 14 or higher) installed on your system. Then, execute the following command:

• Install QLLM CLI globally:

npm install -g qllm

This global installation makes the qllm command readily available in your terminal.

Verify the installation with:

qllm --version

You should see the version number (e.g., 1.8.0) displayed, confirming a successful installation.

4. Configuration

Before diving into the world of AI interactions, configure QLLM CLI with your API keys for the desired LLM providers. QLLM CLI offers flexible configuration management:

Interactive Configuration

Initiate the interactive configuration mode:

qllm configure

This guided process helps you set up API keys and default preferences across several sections:

1. Provider Settings
   • Default Provider
   • Default Model
2. Model Parameters
   • Temperature
   • Max Tokens
   • Top P
   • Frequency Penalty
   • Presence Penalty
3. Other Settings
   • Log Level
   • Custom Prompt Directory
   • Stop Sequence

Command-line Configuration

Usage: qllm configure [options]

Configure QLLM CLI settings.

Options: -l, --list List all configuration settings -s, --set Set a configuration value -g, --get Get a configuration value -h, --help Display help for command

This command allows you to manage configuration settings for the QLLM CLI.

Examples:

qllm configure --set provider=openai
qllm configure --set model=gpt-4

Viewing Current Configuration

Display your current settings at any time:

qllm configure --list

This command shows all current settings, with API keys masked for security.

Configuration File

QLLM CLI stores its configuration in a JSON file located at ~/.qllmrc. While manual editing is possible, using the configure command is recommended for proper formatting and validation.

5. Usage

QLLM CLI offers a variety of commands for interacting with LLMs. Here's an overview of the primary usage patterns:

Running Templates

QLLM CLI allows you to run templates directly. This is now the default behavior when no specific command is provided:

qllm 

For example:

qllm https://raw.githubusercontent.com/quantalogic/qllm/main/prompts/chain_of_tought_leader.yaml

The run command supports various options:

• -p, --provider : Specify the LLM provider (default: openai)
• -m, --model : Choose a specific model
• -t, --max-tokens : Set maximum tokens for the response
• --temperature : Adjust output randomness (0.0 to 1.0)
• -s, --stream: Stream the response in real-time
• -o, --output : Save the response to a file
• `-i, --image