DEV Community: Jigyasa Singh

GhostScript AI — The Haunted Story Engine

Jigyasa Singh — Sat, 31 Oct 2026 16:13:19 +0000

What I Built

I built GhostScript AI — The Haunted Story Engine, an interactive AI-powered storytelling website inspired by the mysterious atmosphere of Halloween and the theme of "Ghost in the Machine."

The idea was to combine spooky storytelling with open-weight AI and create a cinematic experience where users can bring their imagination to life.

Users can:

Choose from different storytelling moods, including Eerie and Mysterious, Cosmic Horror, Dark Fantasy, Playful Halloween and Remembrance.
Enter their own creative prompts.
Generate mysterious stories using an open-weight AI model.
Copy and save their favourite stories.
Access previously saved stories through a personal archive.

The website features a dark cinematic interface, neon-purple effects, animated elements and a responsive layout.

Demo

Live Website:

https://jigyasa-singh1730.github.io/GhostScript-AI-The-Haunted-Story-Engine/

The live demo allows visitors to explore the interface, experiment with story moods and experience the storytelling features.

Code

GitHub Repository:

https://github.com/Jigyasa-singh1730/GhostScript-AI-The-Haunted-Story-Engine

The project is built using:

HTML5
CSS3
JavaScript
Hugging Face Transformers.js
Xenova/flan-t5-small

Feel free to explore the source code and experiment with the project!

How I Built It

I developed GhostScript AI using HTML, CSS and JavaScript, focusing on creating an engaging and interactive user experience.

For the AI component, I integrated Xenova/flan-t5-small using Hugging Face Transformers.js.

The model can perform inference directly in the browser after its model files have been downloaded.

The project also includes a sample-story mode, so visitors can explore the storytelling experience even when the AI model has not been loaded.

I used browser Local Storage to save stories, allowing users to revisit their creations without needing a separate database.

One of my main goals was to make the experience visually engaging while keeping the implementation accessible to beginners and developers interested in open-source AI.

Why Does Open Innovation Matter?

Open innovation matters because it makes experimentation and creativity more accessible.

With open-weight AI models and open-source tools, developers can explore how AI works, experiment with different implementations and build applications without depending entirely on closed AI APIs.

For this project, open-source tools made it possible to integrate an AI model into a browser-based storytelling experience and explore AI-powered creativity.

I believe open innovation gives students and beginners an opportunity to move beyond simply using technology and start building with it.

The Spirit of It

The idea behind GhostScript AI is simple:

Every story deserves to be remembered.

Inspired by the mysterious atmosphere of Halloween and Día de los Muertos, I wanted to create something that connects technology, imagination and remembrance.

A ghost in a machine does not always have to be frightening. Sometimes, it can represent a forgotten story waiting to be told again.

Through this project, I wanted to show that AI can be more than automation — it can also be a tool for creativity and storytelling.

What I Learned

Building this project helped me explore:

Integrating an open-weight AI model into a website.
Working with Transformers.js.
Creating interactive interfaces using JavaScript.
Using Local Storage for persistent browser data.
Designing responsive and visually engaging websites.
Understanding the importance of open-source AI.

This challenge gave me an opportunity to experiment, learn and turn a creative idea into a working project.

Project: GhostScript AI — The Haunted Story Engine

Challenge: Hacktoberfest Open-Source AI Challenge — Week 4

Built with: HTML, CSS, JavaScript and Open-Weight AI

Some stories never sleep. They simply wait for someone to tell them.

FlowEase AI — Intelligent Productivity Assistant

Jigyasa Singh — Sun, 25 Oct 2026 17:47:43 +0000

What I Built

LazyFlow AI — Small Tasks. Big Smiles.

LazyFlow AI is a colorful, browser-based AI productivity assistant designed to make everyday tasks easier, faster, and less stressful.

The idea behind the project is simple: why spend unnecessary time organizing tasks, writing routine emails, or preparing revision notes when AI can help?

LazyFlow AI provides four productivity tools:

Task Tamer: Turns unorganized thoughts and tasks into a structured to-do list.
Email Fairy: Converts rough notes into clear, professional email drafts.
Quick Peek: Summarizes longer text into concise, easy-to-understand information.
Study Buddy: Helps students transform study material into simple revision notes.

The application combines a playful, colorful interface with browser-based AI inference to make productivity feel more approachable.

My goal was to build something that helps students, busy individuals, and anyone dealing with repetitive tasks save effort and focus on what matters most.

Demo

Live Website: LazyFlow AI — GitHub Repository

Source Code: View LazyFlow AI on GitHub

The project is designed with a responsive layout, colorful productivity cards, interactive controls, and a dedicated workspace for AI-generated results.

Note: The link currently provided as the website URL points to the GitHub repository. Replace it with your deployed GitHub Pages URL once the website is published.

Code

GitHub Repository: https://github.com/Jigyasa-singh1730/lazyflow-ai

LazyFlow AI is built with standard web technologies:

HTML5 for the application structure
CSS3 for colorful styling, gradients, animations, and responsive layouts
JavaScript for interactions and AI integration
Hugging Face Transformers.js for running the model in the browser
FLAN-T5 Small as the open-weight text-generation model

The project does not require Ollama, a paid AI API key, or a custom AI backend.

How I Built It

I built LazyFlow AI using HTML, CSS, and JavaScript, integrating the Transformers.js library to run the FLAN-T5 Small model directly in the browser.

Instead of connecting the application to a proprietary AI API, I designed the core workflow around an open-weight model.

Here's how it works:

The user selects one of the four productivity tools.
The user enters text into the workspace.
JavaScript creates a task-specific prompt.
Transformers.js loads the model and runs inference in the browser.
The generated result appears in the interface, where the user can review and copy it.

I also designed a colorful, friendly interface featuring pastel gradients, interactive cards, animated decorations, and responsive layouts.

One important implementation detail is that the model and library are initially loaded from online sources. Although inference runs in the browser, the first download requires internet access.

Why Does Open Innovation Matter?

Open innovation was central to my approach because I wanted to build an AI-powered application without depending on a paid proprietary API.

Using an open-weight model through Transformers.js made it possible to experiment with AI using familiar web technologies while keeping the architecture lightweight.

The main benefits are:

Accessibility: Students and beginner developers can experiment with AI without needing a paid API subscription.
Privacy-conscious processing: User-entered text is processed in the browser rather than sent to a custom application server.
Flexibility: The model and prompts can be modified, and alternative compatible models can be explored.
Lower recurring costs: The application does not require per-request fees from a proprietary text-generation API.
Learning through openness: The project provides a practical way to understand model inference and AI integration in a real web application.

There are trade-offs. The model must initially be downloaded, browser performance varies, and a small model may produce less detailed results than larger models.

For this project, open-weight AI provided an accessible way to build and experiment without introducing the cost and infrastructure of a hosted proprietary API.

My Agent Session

This project uses a browser-based text-generation pipeline rather than a separate autonomous agent framework.

Optional: Add a DevRelay session link here if you recorded and saved your development session.

The Lazy Payoff

LazyFlow AI focuses on reducing the small, repetitive tasks that interrupt a student's day.

Instead of manually arranging scattered tasks, drafting a routine email from scratch, or rewriting study material into shorter notes, users can provide their input and let the AI generate a first draft.

The generated content still needs to be reviewed, but it provides a useful starting point and reduces repetitive effort.

The project is designed around a simple philosophy:

Spend less time on busywork. Keep more time for learning, creating, and relaxing.

Prize Categories

Primary challenge entry: Hacktoberfest Open-Source AI Challenge Week 3 — Slow and Steady.

Additional partner prize categories can be listed here if the project meets their specific requirements.

Built with HTML, CSS, JavaScript, and open-weight AI.

LazyFlow AI — Small tasks. Big smiles. Your own pace.

Side Hustle AI Studio

Jigyasa Singh — Thu, 15 Oct 2026 09:38:43 +0000

This is a submission for the Hacktoberfest Open-Source AI Challenge Week 2: Side Hustle

What I Built

I built SideHustle Studio, a web-based AI-powered platform designed to help aspiring entrepreneurs, students, freelancers, and creators turn their ideas into potential side hustles.

Starting a side hustle can be challenging because people often struggle to find ideas, identify their target audience, and create effective product listings. SideHustle Studio aims to simplify this process by bringing useful tools and resources together in one place.

The goal is to make starting an online side hustle easier, especially for beginners who have limited experience or resources.

Who is it for?

Students looking to explore online income opportunities.
Freelancers and creators who want to showcase their services.
Beginners who need help turning ideas into practical opportunities.
Small business owners who want to improve their product descriptions.

Demo

Live Website:

https://jigyasa-singh1730.github.io/SideHustle-Studio/

Explore the website and try out its available features.

Code

GitHub Repository:

https://github.com/Jigyasa-singh1730/SideHustle-Studio

The source code is available on GitHub for anyone who wants to explore, learn from, or contribute to the project.

How I Built It

I built SideHustle Studio using web technologies:

HTML — for the structure of the website.
CSS — for the styling, layout, and visual design.
JavaScript — for interactive elements and functionality.
Open-source AI — as the foundation for exploring accessible AI-powered side-hustle assistance.

My focus was on creating a beginner-friendly and accessible experience without requiring users to install complicated local AI software.

The project demonstrates how web development and open-source AI can be combined to create useful tools for people exploring entrepreneurship.

Note: The specific AI model or framework should be added here if it is implemented in the project.

Why Does Open Innovation Matter?

Open innovation makes technology more accessible to developers, students, and independent creators.

For a project like SideHustle Studio, open-source technologies offer opportunities to experiment, customize tools, learn from community contributions, and build solutions without depending entirely on closed platforms.

It also encourages collaboration and transparency, allowing other developers to explore the code and help improve the project.

I believe open innovation can help more people turn their ideas into useful digital products, regardless of their experience or resources.

My Agent Session

This project was developed as part of my exploration of AI-powered web development and open-source innovation.

I haven't included an agent session link yet. If I publish a session using DevRelay, I will add it here.

Prize Categories

Challenge: Hacktoberfest Open-Source AI Challenge — Week 2: Side Hustle.

Final Thoughts

Building SideHustle Studio has been an opportunity to explore how web technologies and AI can support people who want to start something of their own.

I look forward to improving the project, learning from community feedback, and making it more useful for aspiring entrepreneurs.

Explore the project:

https://jigyasa-singh1730.github.io/SideHustle-Studio/

If you find the project interesting, check out the repository and share your feedback!

Hacktoberfest #OpenSource #AI #SideHustle #WebDevelopment #JavaScript

TouchGrass AI: Less Scrolling, More Exploring 🌿 | Open-Source AI Challenge

Jigyasa Singh — Fri, 09 Oct 2026 11:33:04 +0000

This is a submission for the Hacktoberfest Open-Source AI Challenge Week 1: Touch Grass

What I Built

I built TouchGrass AI 🌿, an AI-powered outdoor adventure planner designed to encourage people to take a break from their screens and reconnect with nature.

In today's digital world, many of us spend hours in front of screens and sometimes struggle to find simple, enjoyable outdoor activities. TouchGrass AI makes getting outside easier by helping users discover nature-based challenges and plan small outdoor adventures.

Key Features

🌿 AI-Powered Adventure Planning: Generate personalized outdoor activity suggestions.
🌳 Nature Challenges: Discover activities that encourage outdoor exploration.
✅ Challenge Tracking: Mark completed challenges and monitor progress.
📊 Progress Dashboard: Keep track of completed activities.
💾 Local Progress Saving: Store challenge progress in the browser.
🎨 Nature-Inspired Design: A clean, green interface focused on outdoor experiences.
🔄 Fallback Suggestions: Provide predefined activity ideas when AI generation is unavailable.

The project is designed for students, developers, and anyone who wants to balance screen time with real-world experiences.

My goal is simple: Use technology to encourage people to spend less time on screens and more time exploring the world around them.

Demo

🌐 Live Website: TouchGrass AI

Explore the website, discover outdoor challenges, and try the adventure planner.

Code

💻 GitHub Repository: Touch-Grass-Ai

The project is built with HTML, CSS, and JavaScript, with AI integration designed to run in the browser.

The repository contains the source code and project documentation, making it easier for others to explore, learn from, and contribute to the project.

How I Built It

I built TouchGrass AI using HTML5, CSS3, and JavaScript for the frontend, focusing on an interactive experience that makes outdoor exploration simple and engaging.

For the AI component, I used Transformers.js with the Hugging Face model configured in the project: onnx-community/Qwen2.5-0.5B-Instruct.

The goal of this approach is to run a compatible open-source language model directly in the browser, rather than relying entirely on a paid, closed AI API. The application also includes predefined fallback suggestions for situations where AI generation is unavailable.

Technology Stack

HTML5: Page structure and content.
CSS3: Visual design and responsive styling.
JavaScript: Interactions, challenge tracking, and application logic.
Transformers.js: Browser-based AI inference.
Hugging Face: Open model hosting and distribution.
Local Storage: Saving challenge progress in the browser.
GitHub Pages: Hosting the live website.

Building this project helped me explore how open-source AI can be integrated into a practical web application.

Why Does Open Innovation Matter?

Open innovation matters because AI development should be accessible to learners, independent developers, and communities—not just organizations with large budgets.

Using open-source AI tools gives developers an opportunity to experiment, understand how models work, and build useful applications without depending entirely on closed APIs.

For TouchGrass AI, this approach makes experimenting with browser-based AI possible while keeping the application lightweight in its architecture and accessible through a web browser.

Open innovation also encourages collaboration, transparency, learning, and continuous improvement. Developers can study the implementation, suggest improvements, and build upon existing ideas.

As a student exploring AI and web development, this project gave me an opportunity to turn an idea into a working web experience while learning how open-source technologies can support real-world problems.

My Agent Session

I developed this project by working through the frontend structure, styling, JavaScript functionality, and AI integration.

Agent session: Not included in this submission.

Prize Categories

Open-Source AI Application
Browser-Based AI / Web Development

Note: These categories describe the project's focus. Please verify the official challenge partner categories before submitting and replace these with the exact eligible category names.

🌿 TouchGrass AI — Take a break. Step outside. Touch grass.

Thanks for checking out my project! I'd love to hear your feedback and suggestions for making the experience better.