Create Your Own AI Companion With This Step-By-Step Guide

Remember GLaDOS, the iconic AI companion from Portal? That snarky, witty, and (let's be honest) slightly evil AI who guided (and occasionally tried to kill) us through the Aperture Science Enrichment Center? Well, get ready to bring a touch of GLaDOS's unique charm (and passive-aggressive jabs) to your own computer!

This isn't just some pre-programmed chatbot spitting out canned lines. We're talking about a fully interactive, locally-run GLaDOS experience powered by cutting-edge technology, including:

Voice Recognition: Hear your every word (and judge you accordingly).
Natural Language Processing (NLP): Understand the nuances of your language (and use them against you).
Text-to-Speech (TTS): Respond in GLaDOS's signature voice (complete with sarcastic inflections).

Ready to unleash your own digital companion (or nemesis)? Let's get started!

If you would watch the video instead, you can watch here:

Laying the Foundation: Miniconda

Before we summon GLaDOS, we need to ensure your system is ready for this scientific endeavor. While you can venture into the installation process without it, I highly recommend using Miniconda, a fantastic package manager that streamlines dependency management for this project. It's like having your own personal assistant, but for code.

Here's how to set up Miniconda:

Download: Do a quick Google search for "Miniconda Download" and grab the installer for your operating system.
Install: Run the installer and follow the on-screen instructions. It's a straightforward process that should only take a few minutes.

Creating a Virtual Environment for GLaDOS

Now, let's create a dedicated environment for your GLaDOS project. This ensures everything stays organized and isolated, preventing any conflicts with other Python projects or software you might have.

Follow these steps:

Open Anaconda Prompt: Once Miniconda is installed, open the Anaconda Prompt (search for it in your system search).
Create the Environment: Type the following command, making sure to specify Python 3.12:
conda create -n glados python=3.12 -y
content_copyUse code with caution.Bash
Activate the Environment: Once the installation is complete, activate the environment using:
conda activate glados
content_copyUse code with caution.Bash
You'll see "(glados)" appear at the beginning of your command prompt, confirming you're in the right environment.

Acquiring the GLaDOS Project Files

With your environment ready, it's time to retrieve the project files from GitHub.

Navigate to GitHub Repo: Head over to the GitHub repository for the GLaDOS project (link in the description below).
Download the ZIP File: Click on the "Download ZIP" button to download the compressed project files.
Extract the Files: Extract the ZIP file to your desired location. I recommend keeping it organized on your desktop.

Cloning and Installing GLaDOS

Now, let's use git to clone the repository and install GLaDOS.

Open Anaconda Prompt (in the GLaDOS environment): If you closed it, open the Anaconda Prompt again and make sure you're in the "(glados)" environment.
Navigate to Your Desktop:
If you extracted the files to your desktop, type:
cd Desktop
Install git: Search Google for "git for windows" and click on the git-scm link. Download and install the git version for your operating system.
Clone the Repository:
Use the following command, pasting the repo link from the GitHub page:
git clone <repo link>
Navigate to the Project Directory:
Once the cloning process is complete, navigate to the project directory using:
cd glados
Run the Install Script:
Finally, execute the appropriate installation script for your system:
- Windows:
  .\install_windows.bat
- Mac:
  ./install_imac.sh
This script will handle the installation of necessary dependencies, including Whisper for voice recognition and the powerful Llama 38B language model.

Summoning Your AI Companion

With everything set up, you're ready to interact with your own digital GLaDOS. Make sure your microphone is connected and use the following command to start the program:

python ./start_windows.bat

Now, let the science (and the sass) begin!

Personalizing Your GLaDOS Experience

This project is incredibly customizable. You can personalize GLaDOS's responses, change its voice model, and fine-tune her personality to create your own unique AI companion.

Here's a taste of what you can do:

Adjusting Interruptibility: If you're experiencing interruptions due to background noise, you can disable the interrupt feature by changing the "interruptible: true" line to "interruptible: false" in the "glados_config.yaml" file.
Changing Language Models: Experiment with different language models downloaded from Hugging Face. Add them to the "models" folder within the project directory and update the "model_path" in the "glados_config.yaml" file to point to your new model.
Crafting a Unique Personality: Modify the prompts in the configuration file to shape GLaDOS's personality. You can:
- Give her a backstory
- Set her goals
- Provide examples of desired responses

Feeling even more adventurous? You can train your own custom voice model and integrate it into the project for a truly personalized experience.

There you have it! You've successfully navigated the complex world of AI and unleashed your own GLaDOS AI companion. Now, go forth and conduct your own (hopefully less fatal) scientific experiments.