Building Your Own Offline AI Companion

Welcome! This guide will walk you through setting up and customizing your very own offline AI companion. Click through the numbered topics on the left to get started with the essentials, from downloading the files you need to running your AI for the first time.

We’ll guide you step-by-step to make it easy, even if you're not super technical. If you can copy and paste, you’re all set!

Want to see a sneak peek of the end result? Check out the screenshot below to see what we’re aiming for with this AI companion project:

Our goal is to help you create an AI that works entirely offline, and with one crucial feature that many tutorials miss—memory. This AI will remember your conversations, allowing for more personalized and engaging interactions over time.

Ready to get started? Let’s dive into the essentials of setting up your offline AI companion, step by step!

If you have any questions, feel free to reach out to me. This was a fun project, and I’m already using it more than ChatGPT (which I’ve nicknamed "Chatty"). You can email me at

Installing AI - Portable Version Only

If you don't want to go through the whole tutorial and would rather install and run the AI right away, this is for you! These steps will help you quickly set up the AI using a portable version of Python and Ollama. Just follow these simple instructions and enjoy your AI without much hassle. First download the AI Kit

Steps to Install:

• Step 1: Install Ollama (remember that drive E:\ is used for reference only).
• Step 2: Download and install WinPython (Portable Python) and put it in the following folder structure:
  e:\python\python-3.11.9.amd64\
• Step 3: Open a command line (CMD) and run the following command to install the required packages from the requirements.txt file:
  python\python-3.11.9.amd64\python.exe -m pip install -r requirements.txt
• Step 4: Pull a model to use. To get started, I recommend this one:
  ollama pull llama3.2
• Step 5: Edit the app.py , and make sure to edit the chat_model variable to reflect the model you just pulled.
• Step 6: Create a new file called personality.txt. This is where you tell your AI how to behave—be it like a girlfriend or a favorite professor. Save this in the folder where your run_ai.bat is located.
• Step 7: Edit the run_ai.bat as needed, and use this to run the whole setup. Your webpage will be located at: http://127.0.0.1:8000/. Enjoy!
• Step 8: Ensure that all the necessary redistributables are installed. You can download them from this link.

If you find this useful, you can Buy me a red bull! Thanks!