Ai-Personal-Assistant 🖥️✨

A lightweight CLI-based local LLM assistant.

Note: Works only on GNU/Linux.

🌟 Capabilities

• Load and process text/code files
• Browse the web via the assistant
• Support over vision and thinking model
• (More features coming soon!)

Note: Make sure to use a tool_call capable model.

🗒️ To-Do

Nothing here! (for now)

⚙️ Installation

1. Clone the repository:

git clone <repo-url>

2. Create a virtual environment:

python3 -m venv myenv

3. Activate it and install dependencies:

source myenv/bin/activate
pip install -r requirements.txt

4. Enjoy! 🎉

See Additional Setup for optional improvements.

🛠️ Configuration

Basic configuration is in config.py and includes:

• System Prompt
• Model selection
• Ollama endpoint
• Verbose mode
• Streaming mode
• Cosmetic options...

💬 Commands

Command	Description
/bye	Exit the assistant
/stream	Clear context but keep system prompt
/file <absolute/path>	Load a text/code file (.txt, .py, .c, etc.)
/help	Show this help message
/list	List available models
/model <model_name>	Change the current model
/regenerate	Regenerate the last assistant message
/show_config	Show current configuration
/stream True - False	Enable or disable streaming (markdown not supported when True)
/verbose True - False	Enable or disable verbose mode

📝 Additional Setup

To make it easier to run:

1. Open your bash config:

sudo nano ~/.bashrc

2. Add an alias (change paths accordingly):

alias ai='cd /path/to/script_folder && source myenv/bin/activate && python main.py'

3. Reload bash:

source ~/.bashrc

Now you can just type ai to start your assistant. UwU