QLLM: Quantalogic Large Language Model CLI & AI Toolbox 🚀

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, a powerful command-line interface for seamless interaction with Large Language Models (LLMs). QLLM CLI provides a unified platform that supports multiple providers and empowers users with extensive configuration options and features.

Key Highlights:

• Multi-provider support through qllm-lib integration
• Rich, interactive chat experiences with conversation management
• Efficient one-time question answering
• Advanced image input capabilities for visual analysis
• Fine-grained control over model parameters
• Comprehensive configuration system

2. Features

QLLM CLI offers a robust set of features designed for effective AI interaction:

1. 🌐 Multi-provider Support: Seamlessly switch between LLM providers through qllm-lib integration.

2. 💬 Interactive Chat Sessions:

   • Context-aware conversations with history management
   • Real-time streaming responses
   • System message customization

3. ❓ One-time Question Answering: Quick answers for standalone queries with the ask command.

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

   • Local files (Supported formats: jpg, jpeg, png, gif, bmp, webp)
   • URLs pointing to online images
   • Clipboard images
   • Screen captures with display selection

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

   • Temperature (0.0 to 1.0)
   • Max tokens
   • Top P
   • Frequency penalty
   • Presence penalty
   • Stop sequences

6. 📋 Provider Management:

   • List available providers
   • View supported models per provider
   • Configure default provider and model

7. 🔄 Response Handling:

   • Stream responses in real-time
   • Save responses to files
   • Extract specific variables from responses

8. ⚙️ Configuration System:

   • Interactive configuration setup
   • JSON-based configuration storage
   • Environment variable support

3. Installation

To use QLLM CLI, ensure you have Node.js installed on your system. Then install globally via npm:

npm install -g qllm

Verify the installation:

qllm --version

4. Configuration

QLLM CLI provides flexible configuration management through both interactive and command-line interfaces.

Interactive Configuration

Run the interactive configuration wizard:

qllm configure

The wizard guides you through configuring:

1. Provider Settings

   • Default Provider
   • Default Model

2. Model Parameters

   • Temperature (0.0 to 1.0)
   • Max Tokens
   • Top P
   • Frequency Penalty
   • Presence Penalty
   • Stop Sequences

3. Other Settings

   • Log Level
   • Custom Prompt Directory

Command-line Configuration

Set individual configuration values:

qllm configure --set 

View current configuration:

qllm configure --list

Get a specific setting:

qllm configure --get 

Configuration File

Settings are stored in ~/.qllmrc as JSON. While manual editing is possible, using the configure commands is recommended.

5. Usage

QLLM CLI supports three main interaction modes:

1. Direct Questions

qllm ask "What is the capital of France?"

2. Interactive Chat

qllm chat

3. Template-based Execution

qllm run template.yaml

Image Analysis

Include images in your queries:

# Local file
qllm ask "What's in this image?" -i path/to/image.jpg

# URL
qllm ask "Describe this image" -i https://example.com/image.jpg

# Clipboard
qllm ask "Analyze this image" --use-clipboard

# Screenshot
qllm ask "What's on my screen?" --screenshot 1

Response Options

Control output behavior:

# Save to file
qllm ask "Query" -o output.txt

# Disable streaming
qllm ask "Query" --no-stream

# Add system message
qllm ask "Query" --system-message "You are a helpful assistant"

6. Advanced Features

Template-based Execution

QLLM CLI supports running predefined templates:

qllm run template.yaml

Template options:

• -v, --variables: Provide template variables in JSON format
• -ns, --no-stream: Disable response streaming
• -o, --output: Save response to file
• -e, --extract: Extract specific variables from response

Chat Commands

In chat mode, use these commands:

• /help: Show available commands
• /new: Start new conversation
• /save: Save conversation
• /load: Load conversation
• /list: Show conversation history
• /clear: Clear conversation
• /models: List available models
• /providers: List providers
• /options: Show chat options
• /set <option> : Set chat option
• `/image