DEV Community: Sunder Kumar

Founder OS: Turning Notion into an AI Startup, Co-Founder with MCP

Sunder Kumar — Sat, 28 Mar 2026 18:56:50 +0000

This is a submission for the Notion MCP Challenge

Founder OS is an AI powered system that transforms a simple startup idea stored in Notion into a full product plan including competitor research, PRDs, and development roadmaps.

Starting a startup often begins with a simple idea.

But transforming that idea into a clear execution plan requires hours of research, competitor analysis, planning, and documentation.

What if that entire process could happen automatically?

That question led me to build Founder OS.

Demo
What I Built
- The Problem
- The Idea
- What Founder OS Does
Architecture & Technical Design
How I Used Notion MCP
Challenges I Faced
Future Improvements
Conclusion
Useful Links

Demo

In this I demonstrated:

1️⃣ Starting with a raw startup idea

2️⃣ Running AI-powered research

3️⃣ Generating a full Product Requirements Document (PRD)

4️⃣ Creating a roadmap and progress dashboard inside Notion

By the end of the demo, the system automatically transforms a simple idea into a structured startup execution plan.

What I Built

The Problem

Many developers and founders have dozens of startup ideas.

But most ideas never turn into real projects because the planning stage is overwhelming.

Before writing a single line of code, you often need to:

Research competitors
Analyze market gaps
Define features
Write a PRD
Create a roadmap
Track project progress

This process requires switching between multiple tools and dozens of browser tabs.

That friction often kills the momentum.

The Idea

I wanted to build something that acts like a startup operating system.

Instead of manually researching and planning everything, the system should automatically:

Research the market
Analyze competitors
Generate product documentation
Create development timelines
Track progress

And the best place to organize all of this is Notion.

So I built Founder OS.

What Founder OS Does

Founder OS is an AI-powered startup automation framework that transforms a simple idea stored in Notion into a fully structured product plan.

The system integrates:

Notion
AI models (via OpenRouter)
Market research APIs
A custom MCP server

With these components, Founder OS can automatically:

🔍 Perform Market Research

The system searches the web to identify real competitors and market trends.

This helps founders understand the existing ecosystem before building.

📊 Generate Gap Analysis

Founder OS creates structured comparison tables in Notion that show:

competitor features
market gaps
opportunities for differentiation

This helps identify potential blue ocean opportunities.

📅 Build Visual Roadmaps

Founder OS generates Mermaid.js Gantt charts directly inside Notion.

These charts visualize:

project phases
milestones
feature timelines

Turning a vague idea into a clear development roadmap.

📈 Track Project Progress

The system calculates completion percentages based on tasks in the Notion database and generates a progress dashboard.

This provides founders with a real-time overview of how their project is progressing.

Workflow

Founder OS follows a simple but powerful workflow:

Idea → Market Research → PRD Generation → Roadmap Creation → Progress Tracking

Architecture & Technical Design

Founder OS follows a modular service-oriented architecture.

Instead of building one large monolithic script, the system separates responsibilities into dedicated services.

System Architecture

The workflow of Founder OS follows a modular service architecture:

User Idea (Notion Database)
          │
          ▼
MCP Server (Controller)
          │
          ▼
- AI Planning Service
- Research Service
- Notion Integration Service

MCP Server

Acts as the central controller that connects the AI agent to Notion.

It exposes a set of tools that the AI can call.

AI Planning Service

Uses LLMs to:

analyze ideas
generate product plans
structure documentation

Research Service

Connects to external search APIs to gather real-world market data and competitor information.

Notion Service

Handles the complex process of generating structured Notion pages including:

tables
callouts
code blocks
PRD pages

How I Used Notion MCP

The core of this project is the Model Context Protocol (MCP).

Instead of simply sending text to an API, MCP allows the AI to interact with tools and data sources.

I implemented several MCP tools:

`get_startup_ideas`

Reads startup ideas directly from a Notion database.

This allows the AI to understand the founder’s initial concept.

`search_market`

Performs web research to identify:

competitors
market positioning
industry trends

`create_prd_page`

This tool generates a complete Product Requirements Document inside Notion.

It programmatically creates structured blocks including:

PRD sections
feature lists
comparison tables
Mermaid Gantt charts

`generate_status_report`

This tool queries the tasks database and calculates project progress.

It then appends a visual progress report to the idea page.

Using MCP transformed Notion from a static workspace into an active collaborator.

Instead of manually organizing ideas, the AI helps plan and structure the entire project.

Challenges I Faced

Building this system introduced several technical challenges.

Working with Notion Block Structures

Notion pages are composed of nested blocks.

Transforming AI-generated text into properly structured blocks required building a translation layer between AI output and Notion's block API.

Maintaining Cross-Database Relationships

The system connects multiple databases:

Ideas
Tasks
PRDs

Managing relationships between these required careful handling of page IDs and references.

Conclusion

Founder OS changes the role of Notion from a simple note-taking tool into a startup co-founder.

Instead of manually researching, documenting, and organizing ideas, founders can rely on an intelligent system that helps turn ideas into actionable plans.

By combining Notion MCP, AI agents, and real-time research, Founder OS creates a powerful workflow that dramatically reduces the friction between idea and execution.

Useful Links

GitHub Repository: https://github.com/Sunder-Kumar/founder-os

Notion MCP Documentation: https://developers.notion.com/guides/mcp/mcp

Why Building Alone Isn’t Enough: Lessons From a Room Full of Strangers

Sunder Kumar — Tue, 10 Mar 2026 15:32:20 +0000

Walking Into a Room of Ideas and Building an AI Powered Digital Wallet

Have you ever had an idea that seemed too big to execute alone, and yet, by the time you finish, it feels like you’ve created something tangible, meaningful, and alive? That’s exactly what happened when I built my AI Powered Secure Digital Wallet System.

The Solo Builder Mindset

I’ve always been a solo builder. Focused, independent, stubbornly convinced that planning alone could carry me through. I had my notebooks, my carefully organized AI prompts, and my determination to see ideas through.

Collaboration? Optional. Networking? Distracting. Comfort in my own echo chamber felt safer.

But I learned that comfort is a slow way to stay small.

The Idea That Sparked the Project

I wanted to build more than just a digital wallet. I wanted something intelligent, secure, and insightful. Something that could:

Track monthly expenses with detailed charts
Analyze each transaction with AI powered fraud detection
Give users a clear, actionable dashboard to manage their finances

I wanted it to feel personal, trustworthy, and future ready.

Planning the System

I started by defining the architecture.

Backend: Node.js + Express APIs for registration, login, wallet management, and transactions.
Frontend: React + Vite for dashboards, interactive charts, and QR code-based payments.
AI Service: Python + Flask with a Random Forest Classifier to detect suspicious activity in real time.

Every transaction would be analyzed by AI based on patterns like:

Transaction amount vs. sender balance
Time of day
Historical activity patterns

Transactions flagged as high risk were immediately highlighted. Users could trust the system because insights were proactive, not just reactive.

The Solo Build, Reimagined

Even though I was “solo,” I realized I wasn’t truly alone. Designing for users forced me to step into their perspective. Every dashboard tweak, every chart, every alert asked:

“Will this help someone understand their money? Does this inspire trust? Could AI assist without overwhelming?”

This mindset transformed the project from a technical exercise to a product with real world value.

Embracing the Builder Identity

“It’s not about writing perfect code. It’s about turning an idea into reality with the tools you have.”

I spent years chasing the title of “developer,” thinking skill meant everything. But here’s the truth: a builder creates value, ships projects, and learns along the way. Formal credentials are nice, but curiosity, persistence, and courage are enough.

I realized I belonged in tech not because I could code everything from scratch, but because I showed up, directed AI tools, and executed a meaningful idea.

Bringing the Wallet to Life

By the end, I had built an AI Powered Secure Digital Wallet System that was:

User-Friendly: Register, view balance, track transactions, and visualize monthly expenses in charts
Intelligent: AI evaluates transactions in real-time, scoring potential fraud
Interactive: QR code payments pre-fill forms for ease and speed
Secure: Admin dashboard monitors flagged transactions, ensuring oversight and reliability

The AI fraud detection became my favorite feature. Watching patterns emerge, spotting anomalies, and providing instant insights felt like giving users a financial guardian angel.

It wasn’t perfect. Some API calls needed optimization, AI models were trained on demo data, and I still have plenty of tweaks to make. But the wallet exists. It works. And it solves a real problem.

Continuous Growth and Open for Contributors

This project is still evolving. I’m actively working on:

Expanding the AI model to detect more complex fraud patterns
Adding new visualizations for financial insights
Improving authentication and security workflows

I also want to open it up for contributors. If you’re excited about AI, fintech, or building tools that make a difference, this is your chance to collaborate, contribute, and learn alongside the project.

Lessons I Learned Along the Way

Showing up is half the battle. Without starting, nothing exists.
Thinking like a user matters more than solo perfection. Empathy drives impact.
AI amplifies creativity, but vision drives results. Tools are powerful only when guided.
Being a builder isn’t about credentials. Curiosity, persistence, and courage are enough to make something meaningful.

Building this wallet reminded me that tech isn’t about perfection it’s about problem solving, creating value, and daring to share it.

So if you’re waiting for the “perfect moment” or the “right skills,” take this as your sign: show up, dive in, and build. The tools are ready, the ideas are there, and your creation is waiting to come to life.

ContriFlow CLI: An AI-Powered GitHub Contribution Assistant

Sunder Kumar — Sun, 15 Feb 2026 21:09:55 +0000

This is a submission for the GitHub Copilot CLI Challenge

ContriFlow CLI is an AI-powered GitHub contribution assistant that finds beginner-friendly issues, automates setup, generates solution suggestions, and gamifies open-source contributions all from the terminal.

What I Built

ContriFlow CLI - An intelligent command-line tool that democratizes open-source contributions by making it effortless for developers to discover, contribute to, and track open-source projects on GitHub.

The Problem

Many developers want to contribute to open-source projects but struggle with:

Finding beginner-friendly repositories to contribute to
Locating issues that match their skill level
Understanding contribution guidelines
Tracking their contribution progress and streaks
Managing forks, clones, and pull requests from the terminal

The Solution

ContriFlow CLI is a comprehensive GitHub automation tool that handles the entire contribution workflow in one unified interface, making open-source contribution accessible to everyone - from beginners to experienced developers.

Key Features

✨ Smart Issue Discovery

Global search for beginner-friendly issues across all GitHub repos
Filter by programming language, repository stars, and issue labels
Repository-specific issue listing with advanced filtering
Beautiful table-format display with color-coding using cli-table3

🍴 One-Command Setup

Fork repositories directly from the CLI
Clone to organized workspace with conflict detection
Automatic upstream remote configuration
Intelligent contribution guideline detection (3-layer search strategy)

🤖 AI-Powered Solutions

OpenRouter API integration for AI-generated issue solutions
Automatic code patch generation
Smart issue analysis and suggestions

🎮 Gamified Contribution Tracking

Daily contribution challenges with trending repositories
Streak tracking (consecutive contribution days)
XP system and level progression
28+ unique badges (First Step, Streaks, Volume, Collaboration)
Comprehensive contribution dashboard with ASCII visualization

📊 Rich Dashboard

Real-time stats (contributions, PRs, streaks, badges)
Contribution history visualization
Multiple display modes (table, list, ASCII art)
Beautiful box drawing with proper alignment
Progress bars and colorized output

🔐 Secure Authentication

GitHub token-based authentication
Safe credential storage in home directory
Token validation with GitHub API
Support for personal access tokens

📚 Contribution Guidelines Intelligence

Layer 1: GitHub Community Profile API (primary method)
Layer 2: Common repository paths (.github/, docs/, root)
Layer 3: Organization default .github repository
Automatic detection and display of CONTRIBUTING.md

Architecture Highlights

12 Modular Commands:

login/auth - Secure GitHub authentication
search - Repository discovery with filters
issues - Issue finder (global & repo-specific)
fork - One-click repository forking
clone - Smart repository cloning
setup - Complete repository initialization
guide - Contribution guidelines viewer
solve - AI-powered issue solutions
pr - Pull request management
contribute - Gamified contribution tracker
dashboard - Stats and progress visualization
config - Configuration management

Technology Stack:

Node.js + ES Modules
Commander.js for CLI framework
Octokit REST API for GitHub integration
Chalk for beautiful terminal colors
Inquirer.js for interactive prompts
Ora for loading spinners
cli-table3 for formatted tables
Simple-git for Git operations
OpenRouter for AI-assisted issue analysis (optional)

Note: OpenRouter is optional. ContriFlow works fully without AI features and can be used as a standard GitHub
automation CLI.

Demo

╭──────────────────────────────────────────────────────────────╮
│ ╭─╮╭─╮                                                       │
│ ╰─╯╰─╯   ContriFlow CLI v1.0.0                               │
│ █ ▘▝ █  Describe a task or run a command to get started     │
│ ▔▔▔▔                                                      │
│ Use /help for commands. Type exit or quit to leave.          │
╰──────────────────────────────────────────────────────────────╯

📋 Commands: login, search, issues, fork, clone, setup, solve, guide, contribute, pr, dashboard, config
💡 Type /help for command list or /help <command> for details
🔀 Type shift+tab to switch modes (normal/plan/solve)
❌ Type exit or quit to leave

❯ ●

Usage Examples

1. Get Started in 3 Commands:

# Authenticate with GitHub
contriflow login

# Find beginner-friendly issues
contriflow issues github/copilot-cli

# Setup and start contributing
contriflow setup --repo github/copilot-cli --issue 1284

2. Daily Challenge Mode:

# Get today's contribution challenges
contriflow contribute --daily

# Trending repositories with beginner issues
# Across JavaScript, Python, TypeScript, and Go

3. AI-Powered Solutions:

# Get AI-generated solution for an issue
contriflow solve 1284 github/copilot-cli

# Auto-generates patch file with code suggestions

4. Gamified Dashboard:

# View your contribution stats
contriflow dashboard

# Multiple display modes available
contriflow dashboard --ascii      # ASCII art visualization
contriflow dashboard --detailed   # Full history
contriflow dashboard --badges     # All badges
contriflow dashboard --stats      # Statistics only

5. Smart Issue Discovery:

# Global search with filters
contriflow issues --language python --label help-wanted --min-stars 500

# Repository-specific issues
contriflow issues nodejs/node --label bug --state open

# With label filtering
contriflow issues github/copilot-cli --label good-first-issue

Beautiful Output Examples

Command Help System:

$ contriflow --help
Usage: contriflow [options] [command]

🚀 Automate your open-source contributions with ContriFlow

Commands:
  login [options]                      Log in with GitHub
  search [keyword]                     Search repositories
  issues [repo]                        Find issues (global or repo-specific)
  fork [repo]                          Fork a repository
  clone [repo]                         Clone a repository
  guide [repo]                         View contribution guidelines
  solve [issue_number] [repo]          Solve issue with AI
  setup [options]                      Fork and clone
  contribute                           Track contributions & streaks
  pr [issue_number] [repo]             Create pull request
  dashboard                            View stats and progress
  config                               Manage configuration

Beautiful Table Display with cli-table3:

📋 Issues in facebook/react with label: good-first-issue

┌────┬──────────┬──────────────────────────────────────┬──────────────┐
│ #  │ Issue ID │ Title                                │ Labels       │
├────┼──────────┼──────────────────────────────────────┼──────────────┤
│ 1  │ #25432   │ Fix typo in documentation            │ good-first-  │
│    │          │                                      │ issue        │
│ 2  │ #25401   │ Add missing TypeScript types         │ help-wanted, │
│    │          │                                      │ typescript   │
│ 3  │ #25381   │ Update deprecated API usage          │ documentation│
└────┴──────────┴──────────────────────────────────────┴──────────────┘

ASCII Dashboard with Proper Alignment:

╔═══════════════════════════════════════════════════════════╗
║          🎯 ContriFlow Contribution Dashboard 📊         ║
╚═══════════════════════════════════════════════════════════╝

┌─────────────────────────────────────────────────────────┐
│                 ⭐ LEVEL  5 ⭐                         │
│                 📈 XP Progress  45/100                 │
└─────────────────────────────────────────────────────────┘

┌─────────────────────────────────────────────────────────┐
│ 🔥 CURRENT STREAK: 15 days                             │
│ ⭐ LONGEST STREAK: 30 days                             │
│ 🔥 × 15                                                │
└─────────────────────────────────────────────────────────┘

┌─────────────────────────────────────────────────────────┐
│ ✅ CONTRIBUTIONS:   156                                │
│ 📝 ISSUES SOLVED:    28                                │
│ 🔀 PRS CREATED:      12                                │
│ 🎯 TODAY:  5/ 5 ████████████████ 100%                  │
└─────────────────────────────────────────────────────────┘

🏆 BADGES (12 Earned)
  🎯 First Step              🔥 3-Day Streak
  ⭐ 7-Day Streak             👑 30-Day Streak
  💪 10 Contributions         🚀 25 Contributions

Repository & Installation

GitHub: github.com/sunder-kumar/contriflow

Install from NPM:

npm install -g contriflow-cli
contriflow --help

Quick Start

1. Log in to ContriFlow

contriflow login

Interactive login that validates your GitHub token and securely stores your credentials.

Alternative (simpler):

contriflow auth

You'll need a GitHub Personal Access Token with these scopes:

repo - Full control of private repositories
public_repo - Access public repositories
user - User profile data

Create one at: https://github.com/settings/tokens

🔐 Repository Permissions

When creating your GitHub Personal Access Token, configure the following Repository permissions:

Set:

Contents → Read and Write
Pull requests → Read and Write
Issues → Read and Write
Administration → Read and Write (needed for forks)

- Metadata → Read-only (default OK)

⚠️ Most Important

The following permissions are critical for ContriFlow to work correctly:

✅ Contents: Read & Write
✅ Administration: Read & Write

For Development:

git clone https://github.com/sunder-kumar/contriflow.git
cd contriflow-cli
npm install
npm start --help

My Experience with GitHub Copilot CLI

How GitHub Copilot CLI Transformed My Development

GitHub Copilot CLI was instrumental in building ContriFlow CLI. Rather than just using it for code generation, I leveraged Copilot CLI's natural language understanding to solve complex architectural and debugging challenges:

1. Architecture & Design Decisions

@github explain how to structure a Node.js CLI tool with multiple commands

Result: Copilot CLI provided architectural guidance, helping me understand command patterns, middleware setup, and best practices for scalable CLI design that led to the modular 12-command structure.

2. GitHub API Complexity

@github explain the GitHub REST API rate limiting and pagination

Result: Deep understanding of pagination, caching strategies, and efficient API usage patterns - crucial for the issues listing feature that handles thousands of results.

3. Async/Await Debugging

@github why is my Promise not resolving in this async function

Result: Multiple times during development, I had spinner and async issues where Copilot CLI's natural language explanation helped me quickly identify that startSpinner() needed to be awaited - saving debugging time.

4. Data Persistence Strategy

@github show me how to implement a simple JSON-based local database for tracking contributions

Result: Copilot CLI provided patterns for file-based data persistence, which was perfect for a CLI tool that needs to track user contributions locally without external databases.

5. Terminal UI Enhancement

@github how to create beautiful colored tables in the terminal with proper alignment

Result: This led to implementing cli-table3 and chalk, plus creating a padLine() helper function to handle ANSI color codes and emoji width - resulting in the beautiful table displays.

6. Contribution Guidelines Detection

@github how do I find CONTRIBUTING.md in a GitHub repository if it's not in the root

Result: Implemented the intelligent 3-layer search strategy that checks Community Profile API, common paths, and organization defaults - making the tool robust across different repo structures.

7. Gamification Logic

@github how to calculate and track consecutive days without server

Result: Guided me through local timestamp tracking and streak reset logic with proper edge case handling.

8. Error Handling Best Practices

@github what are the best practices for user-friendly error messages in CLI tools

Result: Comprehensive error handling with clear next-step suggestions throughout the codebase.

Impact on Development

Productivity Metrics:

Development time reduced by ~40%
Bug resolution time cut in half
Code quality issues identified early
Better architectural decisions from the start

Learning & Understanding:

Deep GitHub API knowledge
Better async/await patterns
Cleaner error handling strategies
Best practices for CLI user experience
Terminal UI constraints and possibilities

Code Quality Improvements:

More consistent patterns across all 12 commands
Better validation and edge case handling
Improved user experience through clearer messaging
Reduced technical debt

Specific Wins with Copilot CLI

Problem 1: Directory Conflict on Re-Setup
User: contriflow setup --repo github/copilot-cli --issue 1284 (second time)
Error: fatal: destination path already exists and is not an empty directory

Solution via Copilot CLI:

@github how to handle existing directories gracefully in git operations

Led to implementing intelligent directory detection with user options to update, use as-is, or re-clone.

Problem 2: Box Drawing Misalignment
The ASCII dashboard had misaligned borders with emojis and colored text.

Solution via Copilot CLI:

@github why are my box drawing characters misaligned in the terminal

Explained ANSI escape codes and emoji width issues, leading to the padLine() helper function that properly accounts for invisible characters.

Problem 3: Negative Progress Bar Values
Dashboard crashed with "Invalid count value: -200"

Solution via Copilot CLI:

@github how to safely handle negative numbers in progress calculations

Implemented bounds checking and validation throughout the dashboard rendering logic.

Why Copilot CLI Was Essential

Natural Language > Search Engines: Asking "why" questions got better explanations than Stack Overflow
Context Awareness: Providing ContriFlow context made suggestions more relevant
Speed: Instant answers instead of searching documentation
Learning: Clear explanations increased understanding, not just copy-paste solutions
Confidence: AI validation of approaches boosted confidence in implementation

The Real Value

The most valuable aspect of Copilot CLI wasn't code generation - it was problem-solving acceleration. When faced with:

Architectural decisions
Debugging mysterious errors
Understanding complex APIs
Best practice questions
Edge case handling

Having natural language access to that knowledge directly in the terminal transformed my productivity.

Submission Notes

What Makes ContriFlow Special

🎯 Problem Solver: Directly addresses the pain point of finding open-source contribution opportunities

🌍 Inclusive: Designed for contributors of all skill levels with beginner-friendly focus

🤖 AI-Powered: Integration with OpenRouter API for intelligent issue analysis and solutions

🎮 Engaging: Gamification elements (streaks, badges, levels) make contribution fun and rewarding

📈 Comprehensive: Complete workflow from discovery to PR creation - nothing else needed

🚀 Production Ready: Full error handling, validation, testing, and documentation

Challenge Experience

Building ContriFlow CLI was the perfect showcase for GitHub Copilot CLI because:

Real-World Problem: Open-source contribution is a genuine pain point affecting millions
CLI-First Design: Natural fit for a CLI tool that enhances terminal workflows
Copilot Enhanced: Copilot CLI directly improved development speed and quality
Community Benefit: Helps developers everywhere contribute more effectively
Demonstrates Potential: Shows what's possible when AI meets terminal development

Technical Achievements

✅ 15 source files with ~2,000 lines of clean, modular code
✅ 12 CLI commands fully implemented and tested
✅ 3-layer intelligence for contribution guideline detection
✅ Beautiful output with proper box drawing and ANSI handling
✅ Robust error handling with user-friendly messages
✅ Full test coverage with 28/28 tests passing
✅ Comprehensive documentation with 46,000+ words
✅ npm-ready with proper .npmignore and packaging

Thank You

Building ContriFlow CLI with GitHub Copilot CLI was an exciting journey that demonstrated the transformative power of AI-assisted development. The combination of natural language problem-solving and terminal-based productivity created a development experience that was both efficient and enjoyable.

This tool is ready for the community, and I'm excited to see it help developers everywhere contribute to open-source projects!

ContriFlow turns every developer into an open-source contributor, powered by GitHub Copilot CLI.

Storyblok Chatbot 🤖 A Quick Assistant for Support and FAQs

Sunder Kumar — Sat, 28 Jun 2025 20:33:44 +0000

This is a submission for the Storyblok Challenge

What I Built

I built an AI powered chatbot for Storyblok documentation and support. The chatbot answers user questions using content fetched from a Storyblok powered FAQ space, providing instant, context aware help. My goal was to create a seamless, always available support experience for Storyblok users.

The chatbot fetches live FAQ data from Storyblok and uses OpenRouter's Mistral 7B to generate accurate, markdown formatted responses in a conversational style.

Demo

Demo showcases how users can interact with the chatbot by asking various questions about Storyblok documentation. The chatbot retrieves and displays relevant answers in real time, demonstrating its ability to handle a wide range of queries efficiently.
https://storyblok-chatbot-sunder-kumars-projects.vercel.app

Storyblok Space:

Here’s a public API endpoint that returns all published FAQ content from my Storyblok space in JSON format. View Published FAQ Content

Code Repository:

Checkout the repo for detailed instructions and to see all available tools. You can view the full list of tools directly in the Tools Section of the README.
https://github.com/Sunder-Kumar/storyblok_chatbot

Demo Screenshots

Chatbot Interface

How to login in Storyblok?

Who designed you?

Stoyblok Preview

Tech Stack

Next.js (React)
Tailwind CSS
Storyblok (Headless CMS)
OpenRouter (AI API)
Axios

How I Used Storyblok

I used Storyblok as the backend platform to manage the FAQ content seamlessly. The chatbot retrieves FAQs directly from a designated Storyblok space by leveraging the Content Delivery API. This approach guarantees that responses remain consistently accurate and up to date. Additionally, the intuitive interface of the Storyblok editor makes managing and editing content straightforward and efficient.

AI Integration

For the Awesome AI category, I integrated OpenRouter’s API (Mistral model) to generate answers. The AI is prompted with Storyblok FAQ content and instructed to answer only using that information, ensuring accuracy and relevance.

Learnings and Takeaways

This project was a rewarding journey that taught me a great deal about combining cutting edge technologies to solve real world problems. One of the most significant challenges I faced was ensuring that the AI strictly relied on Storyblok content for its responses. Achieving this required careful experimentation with prompt engineering and API configurations.

I’m particularly proud of how quickly I was able to build a functional and valuable tool by leveraging Storyblok’s flexible content management system alongside modern AI APIs. This experience deepened my understanding of integrating headless CMS platforms with AI solutions, as well as the importance of building intuitive, user friendly chat interfaces.

Beyond the technical aspects, this project reinforced the importance of user centric design and the value of creating tools that enhance accessibility and efficiency. By providing instant, reliable support, this chatbot empowers Storyblok users to make the most of the platform, ultimately contributing to a more productive and satisfying experience.

Solo Submission by: Sunder Kumar

Storyblok Headless CMS challenge

Sunder Kumar — Sat, 28 Jun 2025 14:27:44 +0000

dev.to staff for The DEV Team

Jun 11 '25

Join the Storyblok Headless CMS Challenge: $3,000 in Prizes!

#devchallenge #webdev #api #storyblokchallenge

100

Comments 36

4 min read

Flight Stats Hong Kong: Explore Real Time Air Traffic from Hong Kong Airport

Sunder Kumar — Tue, 24 Jun 2025 19:55:50 +0000

🛫 Live Flight Monitoring: A Real Time Dashboard for Aviation Enthusiasts

Hi, I'm Sunder Kumar a Computer Science student and passionate developer. Besides coding, I’ve always been fascinated by how airports handle such complex traffic systems. So I asked myself:

What if I could track and visualize that data in real time?

That’s how Flight Stats HK was born a fully functional web app that tracks live flight statistics using open data from Hong Kong International Airport.

🔗 Live Demo: https://flight-stats-main-sunder-kumars-projects.vercel.app/

🎯 Key Features at a Glance

📅 Date-Based Search – Choose any date to retrieve flight stats.
🔁 Live Data Fetching – Uses AJAX to fetch real-time arrival/departure info.
📊 Dynamic Charts & Histograms – Visualize flights per hour for each direction.
🌍 Top Airports – Discover busiest origins and destinations.
⚡ Performance Snapshot – See total flights, unique destinations, and flight statuses.
📱 Responsive UI – Works smoothly on mobile and desktop.
🚀 Deployed on Vercel – Fast, serverless deployment for global access.

🧭 How the App Works

1️⃣ Select a Date
Start by choosing a date. The app validates your input and prepares to fetch matching flight data.

2️⃣ Retrieve Real-Time Flight Info
Flight statistics are retrieved via a proxy using AJAX (Fetch API) to bypass CORS limitations and access the Hong Kong Airport Open Data API.

3️⃣ View and Analyze

You’ll get:

Total number of departures and arrivals
Status breakdown (e.g., on-time, delayed)
Hourly histogram of flight volume
Rankings of top destination/origin airports All insights update instantly as the data loads.

📊 Dashboard Sections

📈 Flight Summary

Total Departures & Arrivals
Number of active airports
Most common routes

📍 Top Airports
Ranked list of:

Most common destinations (from HK)
Most common origins (to HK)

Includes:

IATA Code
Airport Name
Total flights per airport

📉 Hour-Wise Flight Trends

Histograms for departures & arrivals
Toggle to switch between views
Tooltips and hoverable bars for easy analysis

🧠 Why Airport Data Matters

Airports are incredibly complex systems that rely on data to optimize traffic, avoid delays, and improve passenger experience. With open data becoming more accessible, we can now visualize and understand these systems better than ever.

This project makes real-time aviation data digestible and interactive — useful for aviation geeks, students, and data lovers alike.

⚠️ Current Limitations

Static backend (PHP proxy) used only for CORS workaround
Limited to Hong Kong Airport data only
No map-based visualizations (for now!)

🚀 Planned Future Enhancements

🌍 Support for multiple airports and global flight data
🗺️ Live map of inbound/outbound flights
📱 PWA version for mobile devices
💬 Multi-language support (starting with Chinese & English)
✈️ Integration with airline APIs for more details (e.g., airline logos, aircraft type)

🏁 Final Thoughts
Flight Stats HK was a passion project that allowed me to combine:

API integration
Real-time dashboard design
Data visualization
Clean, responsive UI

Whether you're a traveler, student, or just love air traffic data this app has something to offer.

“Airports are where the world connects. Now, with data you can understand that connection.”

💬 Feel free to try it out, share feedback, or connect with me for collaborations!

📌 Built using:

HTML, CSS, JavaScript, Fetch API, Chart.js, PHP (for API relay)

🚀 Deployed via Vercel

Speech to Text using Assembly AI

Sunder Kumar — Sun, 24 Nov 2024 18:58:50 +0000

This is a submission for the AssemblyAI Challenge : Sophisticated Speech-to-Text.

What I Built

I built a Speech-to-Text Application that showcases the power of Universal-2, AssemblyAI’s latest speech-to-text model. The application:

Supports Multilingual Transcription Users can choose from multiple languages, ensuring global accessibility.
Outputs with Formatting and Timestamps application delivers well-structured transcripts, complete with proper nouns, punctuation, and timestamps.
User-Friendly Interface built using Streamlit, the app features an intuitive frontend for easy navigation and interaction.

Demo

Link to Github Repository

Journey

Incorporating Universal-2:
The application utilizes Universal-2 through AssemblyAI’s robust API. The backend:

Uploads audio files using AssemblyAI's upload endpoint.
Submits transcription requests, including optional parameters like language_code and punctuate.
Polls transcription progress until completion and fetches the final transcript with timestamps, and word-by-word breakdown.

Screenshots

Team Submission:
I worked on this project independently-Sunder Kumar

DEV Community: Sunder Kumar

Founder OS: Turning Notion into an AI Startup, Co-Founder with MCP

Table of Contents

Demo

What I Built

The Problem

The Idea

What Founder OS Does

🔍 Perform Market Research

📊 Generate Gap Analysis

📅 Build Visual Roadmaps

📈 Track Project Progress

Workflow

Architecture & Technical Design

System Architecture

MCP Server

AI Planning Service

Research Service

Notion Service

How I Used Notion MCP

get_startup_ideas

search_market

create_prd_page

generate_status_report

Challenges I Faced

Working with Notion Block Structures

Maintaining Cross-Database Relationships

Conclusion

Useful Links

Why Building Alone Isn’t Enough: Lessons From a Room Full of Strangers

Walking Into a Room of Ideas and Building an AI Powered Digital Wallet

The Solo Builder Mindset

The Idea That Sparked the Project

Planning the System

The Solo Build, Reimagined

Embracing the Builder Identity

Bringing the Wallet to Life

Continuous Growth and Open for Contributors

Lessons I Learned Along the Way

ContriFlow CLI: An AI-Powered GitHub Contribution Assistant

What I Built

The Problem

The Solution

Key Features

Architecture Highlights

Demo

Usage Examples

Beautiful Output Examples

Repository & Installation

Quick Start

1. Log in to ContriFlow

🔐 Repository Permissions

Set:

- Metadata → Read-only (default OK)

⚠️ Most Important

My Experience with GitHub Copilot CLI

How GitHub Copilot CLI Transformed My Development

1. Architecture & Design Decisions

2. GitHub API Complexity

3. Async/Await Debugging

4. Data Persistence Strategy

5. Terminal UI Enhancement

6. Contribution Guidelines Detection

7. Gamification Logic

8. Error Handling Best Practices

Impact on Development

Specific Wins with Copilot CLI

Why Copilot CLI Was Essential

The Real Value

Submission Notes

What Makes ContriFlow Special

Challenge Experience

Technical Achievements

Thank You

Storyblok Chatbot 🤖 A Quick Assistant for Support and FAQs

What I Built

Demo

Storyblok Space:

Code Repository:

Demo Screenshots

`get_startup_ideas`

`search_market`

`create_prd_page`

`generate_status_report`