👻 GhostShell AI: A Haunted Terminal Powered by Local AI
What if your terminal could talk back to you?
For the Hacktoberfest Week 4 challenge, I built GhostShell AI, an experimental haunted-terminal web app that combines a spooky interface with the idea of running AI locally on your own machine.
🔗 GitHub Repository: https://github.com/utkarshydv27/GhostShellAI
👻 What is GhostShell AI?
GhostShell AI turns a regular browser window into a mysterious terminal where you can interact with a fictional digital ghost.
The project explores how local AI can make creative, interactive experiences possible without requiring every conversation to be sent to a cloud-based AI service.
✨ Key Features
- Haunted terminal interface: A spooky, terminal-inspired web experience.
- Local AI integration: Connects to Ollama for generating responses using a locally running language model.
- Ghost whispers: Interact with the ghost through the web interface.
- Session memories: Stores conversation records locally in a JSONL file.
- Offline fallback: Provides fictional responses when the local AI service is unavailable.
- Open-source code: Explore the implementation, run it locally, and contribute improvements.
Note: The offline fallback is simulated; it is not actual AI inference.
🛠️ Tech Stack
- Python
- Flask
- Ollama
- HTML, CSS and JavaScript
- Pytest for automated tests
🚀 How to Run GhostShell AI
You will need Python and Ollama installed on your computer.
Clone the repository:
git clone https://github.com/utkarshydv27/GhostShellAI.git
cd GhostShellAI
Create and activate a virtual environment on Windows:
py -m venv .venv
.\.venv\Scripts\Activate.ps1
Install the dependencies:
python -m pip install -r requirements-dev.txt
Download the default Ollama model and start the application:
ollama pull qwen2.5:1.5b
python app.py
Open the following address in your browser:
To run the automated tests, use:
pytest -q
🔐 Privacy and Safety
GhostShell AI is designed as a fictional terminal experience. It does not execute arbitrary shell commands.
When configured with a local Ollama model, AI generation runs through the local Ollama service. Conversation records are stored in the project's local data directory. Review the source code and README before using the application.
💡 Why I Built This
I wanted to explore a creative side of AI development beyond ordinary chat interfaces. GhostShell AI is an experiment in combining local language models, Python web development, and an immersive user experience.
Projects like this also offer an opportunity to learn about local inference, API integration, frontend design, testing, and open-source collaboration.
🤝 Contributions Welcome
Have an idea to make the ghost smarter, the interface creepier, or the code better?
Check out the repository and contribute improvements:
👉 https://github.com/utkarshydv27/GhostShellAI
Bug reports, feature suggestions, documentation improvements, and code contributions are welcome.
🎃 Final Thoughts
GhostShell AI is a small experiment with a simple idea: AI doesn't always need to live in the cloud to create an interesting experience.
I'm continuing to learn, build, and experiment with open-source AI projects. Feedback and suggestions are welcome!
Challenge: Hacktoberfest Week 4 — Ghost in the Machine
Project: GhostShell AI
Repository: https://github.com/utkarshydv27/GhostShellAI
Top comments (0)