
[Live Demo](https://drive.google.com/file/d/11PcWJr1j021Q3Gc_mWnXIGV4tpSwMe5d/view?usp=sharing)

MedRAG is a fully local, agentic RAG system that lets anyone ask natural-language questions about AI in Healthcare and receive **cited, reasoned answers** grounded in a curated set of 60 expert documents (research papers, market reports, blogs, newsletters). Every response is produced by a 3-agent pipeline — a Decomposer, a Reasoner, and a Critic — and every factual claim is linked to the exact source document that supports it.

![Screenshot-1](/project-files/MedRAG(1).png)

For multi-part questions, Agent 1 automatically decomposes the query into sub-queries (shown as numbered blue chips). The answer body contains **inline [DOC-XXX] badges** — clickable chips that open a document detail popup. The green **Critic Agent — All citations verified** bar confirms Agent 3 checked every claim.

![Screenshot-3](/project-files/MedRAG(3).png)

For questions that span multiple documents, the system retrieves and synthesises across sources. The sidebar lists every document consulted, with its title, publisher, date, and a colour-coded type badge (Market Report, Newsletter, Blog, Research Paper). Citations inside the answer body (`[DOC-039]`, `[DOC-021]`) map directly to the sidebar entries.

![Screenshot-2](/project-files/MedRAG(2).png)

You can read about how this project works in detail over [here](https://github.com/RainboeStrykr/MedRAG/blob/main/DETAILS.md)

MedRAG


![screenshot](/project-files/UniPlan-3D-visualisation.gif)

UniPlan is a modern, full-stack university timetable scheduling application. It uses a **Constraint Satisfaction Problem (CSP)** engine built in Python to automatically resolve complex scheduling variables—like room capacities, lecturer overlap, and student group conflicts—and pairs it with a beautiful, interactive **React** Frontend with advanced 2D/3D visualizations.

## ✨ Features
- **CSP Algorithm Engine**: Python backend wielding Backtracking, Forward Checking, and Advanced Heuristics (MRV, MCV, LCV).
- **Interactive 2D/3D Visualizer**: Watch the algorithm build the schedule in real-time with interactive 3D node graphs and search trees.
- **Dynamic Resource Management**: Add Courses, Professors, and Rooms manually or import directly via CSV.
- **Performance Analytics**: Compare algorithmic node-expansion and execution time natively against baseline Backtracking.
- **Smart Result Generation**: Dynamic timetable mapping displaying fully optimized, clash-free schedules.
- **Comprehensive Documentation**: Built-in documentation explaining the entire system architecture and usage.

[Live Demo](https://drive.google.com/file/d/1c4lwFYpwmgaNzOOLoku4CMiqEaaZATSP/view?usp=sharing)

UniPlan


![Screenshot](/project-files/glyph-pdf.gif)

## Description

Glyph is your all-in-one, browser-based solution for complete PDF management. Designed for simplicity and power, Glyph empowers you to effortlessly merge, split, compress, convert, organise, and secure your PDF documents, all within a private and intuitive environment. From transforming HTML into PDFs to adding watermarks, encrypting sensitive files, or decrypting protected documents, Glyph provides professional-grade tools without requiring downloads or complex installations. Experience unparalleled control over your documents, ensuring privacy and efficiency with every task.

## ✨ Features
- **100% Private** — All processing happens locally in your browser
- **Fast** — No upload/download delays, instant processing
- **Free** — No registration, no limits, no watermarks
- **Responsive** — Works on desktop and mobile

Glyph


![screenshot](/project-files/beamroom.gif)

## *Real-time code-based screen sharing — simple, fast, and secure.*

BeamRoom is a minimal, modern app that lets you **create a room**, share a **unique join code**, and **instantly show your screen** to anyone — no signup required.

## 🚀 Features

* **Join via Code** — Create a room and share a random & unique digit join code
* **Real-Time Screen Sharing** — Low-latency P2P streaming
* **Private & Secure** — Rooms auto-expire; no data stored
* **Cross-Platform** — Works on web & mobile
* **Clean UI/UX** — Minimal, fast, and distraction-free
* **Lightweight Architecture** — No installations for viewers

## 🏗 Tech Stack

* **Next.js v15** - React framework
* **PeerJS** - WebRTC abstraction
* **Tailwind CSS v4 beta** - Styling
* **shadcn/ui** - UI components

Beamroom


![screenshot](/project-files/markdownify.png)

The only markdown editor you will ever need.

## Features

1. **Save Your Files**: You can save your typed docs in local storage

2. **Download Files**: Download your markdown docs as in .md format or in HTML format.

3. **Live Preview**: You can see the live preview of the markdown you are typing for better accuracy and accessiblity.

4. **Light/Dark Mode**: Oh yeah, you got it!

5. **Auto-spellcheck**: Corrects and provides the right grammar and spelling when turned on so you never make another mistake.


Markdownify


![screenshot](/project-files/siteshot-demo.gif)

Generate clean screenshots for any website from its URL.

## Features

- Downloads in PNG format
- Change background color 
- Free and Open source
- Compatible with any website


Siteshot


![screenshot](https://raw.githubusercontent.com/RainboeStrykr/invoswift/refs/heads/main/media/MEDIA.png)

InvoSwift is your smart, fast, and hassle-free solution for creating professional invoices and receipts. Whether you're a freelancer, small business owner, or entrepreneur, InvoSwift helps you stay on top of your finances with just a few taps.

Design invoices and receipts and share them on the go — all from one sleek and easy-to-use platform.

## ⚡️ Features

- **Instant Invoice Generation:** Create and customize invoices in seconds with your company info, items, taxes, and notes.

- **9 Invoice Templates:** Choose from a range of modern, professional templates suited for any business type.

- **Tax Support:** Work globally with support for tax configurations.

- **PDF Sharing Options:** Export invoices and receipts as PDFs.

- **Receipt Generator:** Inbuilt receipt generator with 4 templates.

InvoSwift

<a href="https://drive.google.com/file/d/11PcWJr1j021Q3Gc_mWnXIGV4tpSwMe5d/view?usp=sharing">Live Demo</a>
MedRAG is a fully local, agentic RAG system that lets anyone ask natural-language questions about AI in Healthcare and receive cited, reasoned answers grounded in a curated set of 60 expert documents (research papers, market reports, blogs, newsletters). Every response is produced by a 3-agent pipeline — a Decomposer, a Reasoner, and a Critic — and every factual claim is linked to the exact source document that supports it.
<img src="/project-files/MedRAG(1).png" alt="Screenshot-1">
For multi-part questions, Agent 1 automatically decomposes the query into sub-queries (shown as numbered blue chips). The answer body contains inline [DOC-XXX] badges — clickable chips that open a document detail popup. The green Critic Agent — All citations verified bar confirms Agent 3 checked every claim.
<img src="/project-files/MedRAG(3).png" alt="Screenshot-3">
For questions that span multiple documents, the system retrieves and synthesises across sources. The sidebar lists every document consulted, with its title, publisher, date, and a colour-coded type badge (Market Report, Newsletter, Blog, Research Paper). Citations inside the answer body (<code>[DOC-039]</code>, <code>[DOC-021]</code>) map directly to the sidebar entries.
<img src="/project-files/MedRAG(2).png" alt="Screenshot-2">
You can read about how this project works in detail over <a href="https://github.com/RainboeStrykr/MedRAG/blob/main/DETAILS.md">here</a>