Digital Séance：以開放權重 AI 讓鬼故事重現
Digital Séance 👻 — Bringing Ghost Stories to Life with Open-Weight AI
Digital Séance 是一款開源超自然故事應用，讓用戶輸入回憶或虛構情境、選擇靈體個性與敍事語氣，生成鬼故事。項目以 Qwen2.5-3B-Instruct 作為預定故事模型，透過 Hugging Face Transformers 整合開放權重 AI，並提供替代結局、神秘訊息及個人故事檔案。作者指出，本機推理有助把個人回憶留在用戶自己的裝置，並減少對專有 AI API 的依賴。
This is a submission for the Hacktoberfest Open-Source AI Challenge Week 4: Ghost in the Machine.
What I Built
Digital Séance is an AI-powered supernatural storytelling application that transforms ordinary memories, diary entries, and fictional scenarios into mysterious ghost stories.
Inspired by Halloween and the idea of preserving stories from the past, I wanted to create something that combines supernatural storytelling with open-source artificial intelligence.
Users can enter a memory or a fictional situation, choose a supernatural personality, select a storytelling tone, and generate an atmospheric story.
The experience includes different spirit personalities, alternate endings, mysterious messages, and a personal story archive.
The goal was to create more than another chatbot: an interactive digital space where imagination and open-weight AI come together.
Demo
👻 Live Application: https://digital-s-ance-1.onrender.com
Explore Digital Séance and experience the supernatural storytelling interface.
Code
The project is open source. Explore the repository to see the implementation and contribute to its development.
How I Built It
I built Digital Séance using a modern full-stack architecture.
Technology stack:
- React, Vite and TypeScript for the frontend.
- Tailwind CSS and Framer Motion for the user interface and animations.
- Python and FastAPI for the backend.
- SQLite for story storage.
- Hugging Face Transformers for open-weight AI integration.
- Qwen2.5-3B-Instruct as the intended storytelling model.
The AI component is designed around an open-weight model rather than a closed commercial AI API. The model generates supernatural stories based on user input, selected spirit personalities, and storytelling preferences.
The backend handles model inference and story-generation requests, while the frontend provides an interactive storytelling experience.
Why Does Open Innovation Matter?
Open innovation matters because AI should not be limited to services controlled by a single provider.
Using open-weight models gives developers more control over how AI works, where inference happens, and how applications are customized.
For Digital Séance, this approach creates opportunities for:
- Running AI inference locally.
- Keeping personal memories and stories on the user's own machine.
- Experimenting with different open-weight models.
- Customizing storytelling behaviour.
- Reducing dependence on proprietary AI APIs.
An especially important aspect is privacy. Personal memories can be sensitive, and a locally running model can help keep that information away from third-party inference services.
Open-source tools also make it easier for developers to inspect, modify, and improve the technology behind their applications.
The Spirit of It
Every memory has a story, and sometimes the smallest detail can become the beginning of something mysterious.
Digital Séance explores the idea of bringing forgotten memories and fictional fragments back to life through AI-generated storytelling.
It is a small experiment in combining technology, imagination, and the supernatural.
Some stories are never forgotten. They are just waiting to be summoned.
Prize Categories
- Best Use of Render — For deploying Digital Séance as a live web application using Render.
Built for the Hacktoberfest Open-Source AI Challenge Week 4: Ghost in the Machine. 👻
來源：DEV Community · dev.to