My first MCP server
Learned about the idea of MCP servers, taken the help of AI to build an AI itself and learn a lot about AI agents.
Web Developer, Problem Solver, Detail-Obsessed
|
Hi, I'm Ayush—I Build on the Web
React
Next.js
JavaScript
TypeScript
Java
Spring Boot
Python
Flask
Node.js
MongoDB
PostgreSQL
Supabase
Docker
GitReactNext.jsJavaScriptTypeScriptJavaSpring BootPythonFlaskNode.jsMongoDBPostgreSQLSupabaseDockerGitReactNext.jsJavaScriptTypeScriptJavaSpring BootPythonFlaskNode.jsMongoDBPostgreSQLSupabaseDockerGitReactNext.jsJavaScriptTypeScriptJavaSpring BootPythonFlaskNode.jsMongoDBPostgreSQLSupabaseDockerGitI design scalable Apps
My tech stack
I constantly try to improve
|
Currently building a JS Animation library
The Inside Scoop
Do you want to start a project together?
A small selection of
Recent projects
Discuss - A Community Platform
June 2025 - July 2025
Discuss is a full-stack community platform built for open discussions and knowledge sharing. Whether it's tech, education, hobbies, or support groups, Discuss provides a clean, modern interface for users to post, comment, and connect.
Next.jsTailwindCSSTypescriptThree.jsFramer Motion
X Agent-An AI MCP Agent
Feb 2025 - March 2025
A Node.js agent demonstrating the power of combining Generative AI for content creation with the Model Context Protocol (MCP) for reliable tool execution. This project provides an interactive CLI that can automatically compose and post tweets on twitter.
Node.jsMCP/StreamsLLM
AI Image Improver
April 2025
AI Image Enhancer is a web-based tool that uses AI to upscale and enhance low-resolution images. Built using React, Tailwind CSS, and AI-powered image enhancement APIs
ReactTailwindCSSTypescript
Project stories
Discuss
It started as a simple idea — a place where people could share thoughts in real-time. I built it with Next.js, WebSockets, and Prisma, crafting live threads and instant comments.
KhaataBook
Managing daily expenses was always a pain, so I set out to simplify it. I created KhaataBook, a personal expense manager with secure, per-user accounts. Later, I added encryption so sensitive files stayed private, and a shareable link system for quick collaboration. It grew from a basic tracker into a secure, user-friendly finance tool.
IPL Predictor
As a cricket fan and data geek, I wanted to bring stats to life. I trained a Logistic Regression model on years of IPL data, predicting match winners with surprising accuracy. The model sparked debates, laughter, and bragging rights among my friends — and taught me a ton about machine learning.
PokeDex
Every Pokémon fan knows the joy of flipping through a Pokédex. I decided to recreate that magic with a modern twist — a web app featuring detailed stats, evolutions, and attack lists. With a smooth UI and fast search, it became both a nostalgic trip and a useful companion for fans.1
My first MCP server
Learned about the idea of MCP servers, taken the help of AI to build an AI itself and learn a lot about AI agents.
Discuss
It started as a simple idea — a place where people could share thoughts in real-time. I built it with Next.js, WebSockets, and Prisma, crafting live threads and instant comments.
KhaataBook
Managing daily expenses was always a pain, so I set out to simplify it. I created KhaataBook, a personal expense manager with secure, per-user accounts. Later, I added encryption so sensitive files stayed private, and a shareable link system for quick collaboration. It grew from a basic tracker into a secure, user-friendly finance tool.
IPL Predictor
As a cricket fan and data geek, I wanted to bring stats to life. I trained a Logistic Regression model on years of IPL data, predicting match winners with surprising accuracy. The model sparked debates, laughter, and bragging rights among my friends — and taught me a ton about machine learning.
PokeDex
Every Pokémon fan knows the joy of flipping through a Pokédex. I decided to recreate that magic with a modern twist — a web app featuring detailed stats, evolutions, and attack lists. With a smooth UI and fast search, it became both a nostalgic trip and a useful companion for fans.1
My first MCP server
Learned about the idea of MCP servers, taken the help of AI to build an AI itself and learn a lot about AI agents.
Discuss
It started as a simple idea — a place where people could share thoughts in real-time. I built it with Next.js, WebSockets, and Prisma, crafting live threads and instant comments.
KhaataBook
Managing daily expenses was always a pain, so I set out to simplify it. I created KhaataBook, a personal expense manager with secure, per-user accounts. Later, I added encryption so sensitive files stayed private, and a shareable link system for quick collaboration. It grew from a basic tracker into a secure, user-friendly finance tool.
IPL Predictor
As a cricket fan and data geek, I wanted to bring stats to life. I trained a Logistic Regression model on years of IPL data, predicting match winners with surprising accuracy. The model sparked debates, laughter, and bragging rights among my friends — and taught me a ton about machine learning.
PokeDex
Every Pokémon fan knows the joy of flipping through a Pokédex. I decided to recreate that magic with a modern twist — a web app featuring detailed stats, evolutions, and attack lists. With a smooth UI and fast search, it became both a nostalgic trip and a useful companion for fans.1
My first MCP server
Learned about the idea of MCP servers, taken the help of AI to build an AI itself and learn a lot about AI agents.
Discuss
It started as a simple idea — a place where people could share thoughts in real-time. I built it with Next.js, WebSockets, and Prisma, crafting live threads and instant comments.
KhaataBook
Managing daily expenses was always a pain, so I set out to simplify it. I created KhaataBook, a personal expense manager with secure, per-user accounts. Later, I added encryption so sensitive files stayed private, and a shareable link system for quick collaboration. It grew from a basic tracker into a secure, user-friendly finance tool.
IPL Predictor
As a cricket fan and data geek, I wanted to bring stats to life. I trained a Logistic Regression model on years of IPL data, predicting match winners with surprising accuracy. The model sparked debates, laughter, and bragging rights among my friends — and taught me a ton about machine learning.
PokeDex
Every Pokémon fan knows the joy of flipping through a Pokédex. I decided to recreate that magic with a modern twist — a web app featuring detailed stats, evolutions, and attack lists. With a smooth UI and fast search, it became both a nostalgic trip and a useful companion for fans.1
My first MCP server
Learned about the idea of MCP servers, taken the help of AI to build an AI itself and learn a lot about AI agents.
Discuss
It started as a simple idea — a place where people could share thoughts in real-time. I built it with Next.js, WebSockets, and Prisma, crafting live threads and instant comments.
KhaataBook
Managing daily expenses was always a pain, so I set out to simplify it. I created KhaataBook, a personal expense manager with secure, per-user accounts. Later, I added encryption so sensitive files stayed private, and a shareable link system for quick collaboration. It grew from a basic tracker into a secure, user-friendly finance tool.
IPL Predictor
As a cricket fan and data geek, I wanted to bring stats to life. I trained a Logistic Regression model on years of IPL data, predicting match winners with surprising accuracy. The model sparked debates, laughter, and bragging rights among my friends — and taught me a ton about machine learning.
PokeDex
Every Pokémon fan knows the joy of flipping through a Pokédex. I decided to recreate that magic with a modern twist — a web app featuring detailed stats, evolutions, and attack lists. With a smooth UI and fast search, it became both a nostalgic trip and a useful companion for fans.1
My first MCP server
Learned about the idea of MCP servers, taken the help of AI to build an AI itself and learn a lot about AI agents.
Discuss
It started as a simple idea — a place where people could share thoughts in real-time. I built it with Next.js, WebSockets, and Prisma, crafting live threads and instant comments.
KhaataBook
Managing daily expenses was always a pain, so I set out to simplify it. I created KhaataBook, a personal expense manager with secure, per-user accounts. Later, I added encryption so sensitive files stayed private, and a shareable link system for quick collaboration. It grew from a basic tracker into a secure, user-friendly finance tool.
IPL Predictor
As a cricket fan and data geek, I wanted to bring stats to life. I trained a Logistic Regression model on years of IPL data, predicting match winners with surprising accuracy. The model sparked debates, laughter, and bragging rights among my friends — and taught me a ton about machine learning.
PokeDex
Every Pokémon fan knows the joy of flipping through a Pokédex. I decided to recreate that magic with a modern twist — a web app featuring detailed stats, evolutions, and attack lists. With a smooth UI and fast search, it became both a nostalgic trip and a useful companion for fans.1
Roles
Full-Stack Web Developer
Developed and maintained web applications using a variety of technologies, including Next.js, Node.js, and Databases.
Role #01
Backend Developer
Designed and developed server-side logic and APIs using Node.js and Express.
Role #02
Problem Solving
Utilized analytical skills to troubleshoot and resolve complex issues in software applications.
Role #03
AI/ML
Explored AI/ML concepts and implemented basic models using Python and relevant libraries.
Role #04
Always learning, always refining.
My Approach
Phase 1
Understanding the Problem
Break down the problem into modular components by identifying key tasks and dependencies. This helps in creating a clear roadmap for development.
Phase 2
Design and Build
Design and iterate quickly. Build clickable prototypes or POCs, validate assumptions, and refine based on feedback.
Phase 3
Test, Optimize & Deploy
Ship confidently: harden, test, and document. Monitor performance, and apply polish for a stable launch.