GhostMemory AI — Where Memories Never Die 👻

Title: GhostMemory AI — Eternal Memories in AI <br>

This project is part of the Hacktoberfest Open-Source AI Challenge Week 4: Ghost in the Machine. It is an AI-powered storytelling experience that explores the notion that memories never truly vanish—they can resurface as stories, fragments, and echoes from the past. The aim was to investigate the intersection of memory, imagination, and artificial intelligence. <br>

The application is built on the idea that an old memory could be transformed into a mysterious tale or an unfinished diary entry could be reimagined into a haunting narrative. It is designed with a dark, atmospheric interface and includes features for crafting eerie stories, safeguarding memories, reimagining diary entries, and exploring saved stories. <br>

Key Features include: <br>

- Create Story: Convert a premise or concept into a paranormal mystery. <br>

- Memory Keeper: Delve into the idea of preserving significant memories. <br>

- Diary Reimaginer: Convert diary entries and forgotten moments into fresh narratives. <br>

- My Stories: Compile storytelling experiences in one place. <br>

- Cloud AI Integration: Connect the app to Cloudflare Workers AI via server-side functions. <br>

The demo of the application can be accessed at https://ghostmemory-ai.netlify.app/. The source code is available on GitHub at https://github.com/priyarani22023-max/GhostMemory-AI. <br>

The development of GhostMemory AI began with Google AI Studio, which was used to build and refine the application. It was then linked to GitHub and deployed on Netlify. The AI architecture planned for the project involves using Cloudflare Workers AI for inference through an open-weight model. Server-side requests are managed by Netlify Functions to keep API credentials out of the frontend. <br>

The workflow is as follows: a user inputs a story premise or memory. The request is sent to a Netlify Function, which then calls the configured Cloudflare Workers AI model. The generated output is returned to the frontend and displayed to the user. The application is set up to use environment variables for the Cloudflare Account ID, API token, and model ID, instead of hardcoding sensitive information into the source code. <br>

This serverless architecture allows the project to be accessed online without requiring users to install a local inference engine. The choice of this approach enables experimentation with creative storytelling while learning about model integration, serverless functions, deployment, and secure credential management. It also simplifies the project for community members to inspect, improve, and extend.