DEV Community: Tanvi Sharma

GhostStory AI 👻: Creating Spooky Stories with Local Open-Source AI

Tanvi Sharma — Sat, 31 Oct 2026 14:26:54 +0000

GhostStory AI 👻

For Hacktoberfest 2026 Week 4, I built GhostStory AI — a spooky story generator powered by local open-source AI.

👻 What does it do?

Creates original ghost stories.
Lets users choose a main character.
Generates stories in different settings.
Offers scary, mysterious, and emotional moods.

🛠️ Tech Stack

HTML
CSS
JavaScript
Ollama
Qwen 2.5 (1.5B)

🧠 Why Open Innovation Matters

GhostStory AI runs Qwen locally through Ollama. This gives users more control over their prompts and reduces dependence on hosted AI services. Open-weight models make AI experimentation more accessible to beginners and developers.

💻 Source Code

07Tanvi / ghoststory-ai-tanvi26

GhostStory AI 👻

A spooky story generator powered by open-source AI.

Features

Generate original ghost stories.
Choose characters and spooky settings.
Select scary, mysterious, or emotional moods.
Runs locally using Ollama.

Tech Stack

HTML, CSS, JavaScript
Ollama
Qwen 2.5 (1.5B)

Why Open Innovation?

GhostStory AI uses an open-weight model running locally. This gives users more control over their prompts and stories, reduces dependence on hosted AI services, and makes experimentation accessible.

Run Locally

Install Ollama.
Download qwen2.5:1.5b.
Run python -m http.server 8003 in the project folder.
Open http://localhost:8003.

Demo

A demo video link will be added here.

View on GitHub

I enjoyed building a small creative project that brings a little mystery and imagination to AI.

devchallenge #hf26challenge #opensource #ai #hacktoberfest

LazyMate AI ✨: Making Repetitive Tasks Easier with Local AI | Hacktoberfest 2026 Week 3

Tanvi Sharma — Fri, 23 Oct 2026 04:32:37 +0000

LazyMate AI ✨ — Making Repetitive Tasks Easier with Local AI

Ever put off a simple task because it feels too overwhelming to start?

For the Hacktoberfest 2026 Week 3 “Slow and Steady” challenge, I built LazyMate AI — a small productivity assistant that turns boring tasks into simple steps, a quick checklist, and a time-saving suggestion.

🌱 What it does

Breaks a task into manageable steps.
Creates a simple checklist.
Suggests a shortcut to save time.
Helps you get started with less stress.

🧠 Why open-source AI?

LazyMate AI uses the open-weight Qwen 2.5 model through Ollama on my computer. Open models make AI experimentation more accessible and give developers the freedom to build and learn locally.

🛠️ Built with

HTML, CSS, and JavaScript
Ollama
Qwen 2.5 (1.5B)

💻 Source code

https://github.com/07Tanvi/lazymate-ai-tanvi26

Building this project reminded me that AI doesn't always need to do something huge. Sometimes, helping someone take the first small step is enough.

devchallenge #hf26challenge #opensource #ai #productivity

SellSmart AI: Product Listings Made Easier with Local AI 🛍️

Tanvi Sharma — Tue, 13 Oct 2026 16:18:19 +0000

I Built SellSmart AI: Product Listings Made Easier with Local AI 🛍️

For Week 2 of the Hacktoberfest 2026 DEV Open-Source AI Challenge, I built SellSmart AI — a simple tool that helps small businesses and online sellers create product marketing content.

💡 What does it do?

SellSmart AI lets users enter a product name, add product details, and choose a selling platform. It then generates a catchy title, product description, and promotional caption.

🛠️ Tech Stack

HTML
CSS
JavaScript
Ollama
Qwen 2.5 1.5B

🌱 Why Open-Source AI?

I used an open-weight model running locally through Ollama. This approach gives users more control over their data and reduces dependence on paid, hosted AI APIs.

As a beginner, this project helped me learn how a web interface can connect to a local AI model and turn user input into useful content.

🎯 Who is it for?

SellSmart AI is designed for small business owners, freelancers, and online sellers who want to create product listings more quickly.

🔗 Project

GitHub repository: https://github.com/07Tanvi/sellsmart-ai-tanvi26

🚀 What's Next?

I'd like to improve the interface, test more product types, and gather feedback from real sellers.

Built for Hacktoberfest 2026 — Week 2: Side Hustle.

devchallenge #hf26challenge #opensource #ai

I Built TouchGrass AI: A Little Less Scrolling, A Little More Outside

Tanvi Sharma — Sat, 10 Oct 2026 12:13:50 +0000

🌿 TouchGrass AI — Find Your Outside

I built TouchGrass AI for the Hacktoberfest 2026 Week 1 “Touch Grass” challenge.

It is a simple web app that suggests outdoor activities based on a user's mood, available time, and preferred outdoor setting.

How it works

The app uses HTML, CSS, and JavaScript for its interface and connects to a locally running Qwen 2.5 1.5B open-weight model through Ollama to generate suggestions.

Why I built it

I wanted to make a small tool that encourages people to take screen breaks, enjoy nature, and try simple outdoor activities.

Why open-source AI?

Running an open-weight model locally helped me learn how AI can be integrated into a web app without relying on a paid hosted AI API.

Project link :https://github.com/07Tanvi/touchgrass-ai

This is my beginner project, and I look forward to improving it further!

#devchallenge #hf26challenge