jenix 0.1.3

An agentic coding assistant using local LLMs from Ollama

Jenix

Jenix is an agentic coding assistant that works directly in your terminal, leveraging local LLMs from Ollama. It simplifies coding tasks, provides intelligent suggestions.

---

🚀 Features

• Interactive Chat Mode: Have conversations with AI models for coding assistance.
• Agent Mode: Generate, edit, and manage code files with ease.
• Advanced AI Support:
  • Local Ollama models
• Enhanced Tools: File operations, code analysis, and terminal commands.
• Cross-Platform: Works on Windows, macOS, and Linux.

---

🛠️ Prerequisites

Before using Jenix, ensure the following:

1. Python: Version 3.6 or higher.
2. Ollama: Installed and running. Download Ollama.
3. Language Models: Pull one or more models into Ollama (e.g., codellama:13b, mistral:7b).

---

📦 Installation

1. Clone the Repository

git clone https://github.com/Abhijeetsingh610/Jenix.git
cd Jenix

2. Choose one of the installation methods:

   Option 1: Pip installation

   pip install -e .
   

   Option 2: Windows batch file

   install.bat
   

   Option 3: PowerShell users (recommended for PowerShell)

   .\install_powershell.bat
   

   This creates a PowerShell module that allows you to use the Jenix command directly in PowerShell.

Usage

Basic Commands

Jenix --model MODEL_NAME --chat       # Start a chat session
jenix --model MODEL_NAME --agent      # Start an agent for coding tasks
jenix --claude-agent                  # Start Claude Agent (requires API key)
jenix --jenix-pro-agent            # Start Jenix Pro Agent (Gemini integration)
jenix --list-models                   # Show available models
jenix --no-color                      # Disable colored output
jenix help                            # Show help information

Windows Users

CMD Users: Run run_jenix.bat for a convenient menu interface.

PowerShell Users: After running install_powershell.bat, you can use jenix commands directly in PowerShell. Alternatively, use .\Use-Jenix.ps1 with arguments (the simplest approach).

Chat Mode

Chat mode allows you to have a conversation with the LLM:

jenix --model llama2:7b --chat

Chat Commands

While in chat mode, you can use these special commands:

• exit or quit - End the session
• clear - Clear chat history
• save - Save conversation to file
• help - Show available commands

Agent Mode

Agent mode helps you generate code based on your requirements:

jenix --model codellama:13b --agent

How Agent Mode Works

1. Describe what you want to build
2. The agent will generate code and list files it plans to create
3. You can view the full LLM response if needed
4. You'll be asked to confirm each file creation
5. The agent checks for existing files and handles directories automatically
6. After file creation, you can run executable files directly

Installation

Basic Installation

# Clone the repository
git clone https://github.com/Abhijeetsingh610/Jenix.git
cd jenix-pro-agent

# Install the package
pip install -e .

Windows-specific Installation

For Windows users, you'll need the pyreadline3 package:

pip install -e ".[windows]"

Usage

Run the agent with a local Llama model:

jenix --model llama3:8b --jenix-pro-agent

Requirements

• Python 3.8 or higher
• Ollama (for local models)

## Using Ollama Models

To use local Ollama models, make sure Ollama is installed and running on your system, then:

```bash
jenix --model llama3:8b --jenix-pro-agent --use-ollama

You can also set the jenix_OLLAMA_MODEL environment variable to specify the model:

set Jenix_OLLAMA_MODEL=llama3:8b
jenix --jenix-pro-agent --use-ollama

Chat Commands

• /exit or /quit - Exit the chat session
• /help - Show help information

Troubleshooting

Windows-specific Issues

• If you encounter ModuleNotFoundError: No module named 'readline', run the Windows fix script as described above.

Ollama Connection Issues

• Make sure Ollama is running locally at http://localhost:11434
• Check that the requested model is available in your Ollama installation

Examples

Listing Available Models

jenix --list-models

Starting Chat Mode

jenix --model mistral:7b --chat

Using Agent Mode to Create and Run a Python Script

jenix --model codellama:13b --agent

Then describe your coding task when prompted, such as: "Create a simple weather API client in Python"

### Automatic Model Selection

If you don't specify a model, Jenix will automatically use the first available model in your Ollama installation:

jenix --chat jenix --agent

## License

MIT