This is a submission for the Hacktoberfest Open-Source AI Challenge Week 4: Ghost in the Machine
What I Built
PhantomAI is a spooky AI-powered coding companion that brings a little Halloween magic to programming.
Inspired by the theme Ghost in the Machine, I built an interactive AI assistant that feels like communicating with a friendly phantom from the digital world.
With PhantomAI, users can:
- Ask coding questions and get explanations.
- Understand programming concepts in beginner-friendly language.
- Identify bugs and receive suggestions for fixes.
- Generate spooky Halloween stories.
- Explore fun, haunted coding challenges.
The goal was to combine the Halloween spirit with practical AI-powered programming assistance.
Demo
🔗 GitHub Repository:
https://github.com/001abhimanyu/PhantomAI-The-Haunted-Code-Companion
PhantomAI runs locally on your machine using Ollama and an open-weight AI model.
To experience it, clone the repository, install Ollama, download the model, and launch the website.
Code
👨💻 GitHub Repository:
https://github.com/001abhimanyu/PhantomAI-The-Haunted-Code-Companion
The project is built using:
- HTML
- CSS
- JavaScript
- Ollama
- Qwen 2.5 3B (open-weight model)
How I Built It
I wanted to create something that was both fun and useful while keeping open-source AI at its core.
I used Qwen 2.5 3B through Ollama for local AI inference.
The frontend is built using HTML, CSS, and JavaScript, with a Halloween-inspired interface featuring dark visuals, glowing accents, and a haunted chat experience.
The AI integration communicates with Ollama's local API, allowing users to interact with the model directly from the browser.
The setup is simple:
- Install Ollama.
- Download the model using
ollama run qwen2.5:3b. - Launch the PhantomAI frontend.
- Start chatting with the Phantom!
No API key or paid AI subscription is required.
Why Does Open Innovation Matter?
Open innovation is the heart of PhantomAI.
Using an open-weight model and local inference made it possible to build an AI experience without depending on a closed, proprietary API.
Here's what this approach makes possible:
- Privacy: Conversations and code can remain on the user's local machine.
- Accessibility: No paid API subscription is required.
- Freedom: Users can experiment with different open-weight models.
- Independence: The project isn't locked into a single AI provider.
- Offline potential: After downloading the model, inference can work without an internet connection.
For a coding assistant, privacy matters. Developers may work with sensitive code, and local inference gives them greater control over their data.
Open-source AI makes experimentation, customization, and learning accessible to everyone.
My Agent Session
I used PhantomAI to experiment with programming explanations, spooky storytelling, and coding assistance.
Optional section: If you have a DevRelay agent session link, add it here.
The Spirit of It
Halloween has always been a time for mysterious stories, strange encounters, and things hiding in the shadows.
I wanted to bring that atmosphere into the world of programming.
What if your coding assistant wasn't just another chatbot? What if it felt like a ghost living inside your terminal, ready to help you solve bugs and explain forgotten code?
That idea became PhantomAI.
The project is a reminder that AI doesn't always need to live in the cloud. Sometimes, the most interesting intelligence is the one hiding inside your own machine.
Prize Categories
Open-source AI · Local AI Inference · Creative AI Applications
Built with curiosity, open-source AI, and a little Halloween spirit. 👻
Thanks for participating in Hacktoberfest!
Top comments (0)