Case Studies

How I Build SaaS Products

Real projects with technical challenges, founder insights, and lessons learned. See how I think through problems and ship fast.

kanbi - Action Board

7 days

Tool that converts chaotic brainstorming into organized Kanban boards.

Why I Built This

I noticed founders waste hours organizing scattered ideas from Slack, emails, and meetings. They need to move fast, but manual task organization kills momentum. I wanted to prove AI could turn messy text into actionable boards instantly.

The Solution

Built an AI powered Kanban board that converts unstructured text into organized tasks with automatic categorization and priority suggestions.

Technical Challenges

Parsing unstructured text reliably was hard AI sometimes misinterpreted context. Solved by adding prompt engineering with examples and validation rules. Cookie based storage required careful state management to prevent data loss on refresh.

Founder Learning

Founders don't need perfect they need fast. This tool proves you can ship AI features in days, not months. The key is scoping ruthlessly: one core workflow, done well.

Results

Fully functional SaaS with AI integration, real time updates, drag and drop interface, and team collaboration. Deployed on Vercel with cookie storage.

Next.js
TypeScript
AI Integration
Real time
Drag & Drop
Built in 7 days

Clario - Tool

21 days

Summarization tool that converts long form content into concise summaries.

Why I Built This

Founders drown in content investor decks, competitor research, industry reports. Reading everything takes hours they don't have. I built this to prove AI can compress information without losing what matters.

The Solution

Created an AI summarization tool that processes text, PDFs, and URLs into concise summaries with adjustable length and tone options.

Technical Challenges

PDF parsing broke with complex layouts and images. Fixed by implementing fallback extraction methods and text cleaning. URL scraping hit rate limits added caching and retry logic with exponential backoff.

Founder Learning

Content tools need to handle messy real world inputs. Users will throw anything at it broken PDFs, paywalled URLs, weird formatting. Build for chaos, not perfect test cases.

Results

Production ready app with AI API integration, file upload handling, and responsive design. Shows ability to build content processing tools with modern UI.

Next.js
TypeScript
Content
Summarizer
Text Analysis
Built in 21 days

Subsight - Tracker

14 days

Privacy first subscription tracker PWA with local storage. Features PWA capabilities.

Why I Built This

Founders burn cash on forgotten subscriptions $50 here, $100 there. Existing trackers require accounts and store sensitive financial data. I wanted to prove you can build useful tools that respect privacy completely.

The Solution

Built a Progressive Web App that stores all data locally in the browser. Includes renewal reminders, spending charts, and works offline.

Technical Challenges

Service workers for offline support were tricky cache invalidation caused stale data bugs. Solved with versioned caching strategy. Local storage has size limits, so implemented data compression for users with 50+ subscriptions.

Founder Learning

Not every SaaS needs a backend. Local first apps are faster, cheaper to run, and users trust them more. Perfect for MVPs where you're testing demand before building infrastructure.

Results

Fully functional PWA with local first architecture, service workers for offline support, and installable on mobile devices. No backend required.

Next.js
TypeScript
PWA
Privacy
Tracker
Built in 14 days

Reckoner - Calculator

14 days

Financial runway calculator for startups. Features financial modeling capabilities.

Why I Built This

Every founder I talked to tracked runway in messy spreadsheets. They'd miscalculate burn rate and run out of money unexpectedly. I built this to give founders a clear, visual answer to 'how long do we have?'

The Solution

Developed a runway calculator with visual charts, scenario planning, and burn rate analysis. Helps founders see their financial runway clearly.

Technical Challenges

Real time chart updates caused performance issues with large datasets. Optimized by debouncing calculations and using React.memo for chart components. Export functionality required generating clean PDFs from dynamic charts solved with html2canvas and jsPDF.

Founder Learning

Financial tools must be dead simple. Founders are stressed about money complex UIs make it worse. Big numbers, clear charts, instant feedback. That's what converts anxiety into action.

Results

Production ready financial tool with real time calculations, interactive charts, and export functionality. Demonstrates handling complex math and data visualization.

Next.js
TypeScript
Finance
Charts
Analytics
Built in 14 days

Pitchcraft

7 days

Pitch deck generator for startups with document generation capabilities.

Why I Built This

Founders spend weeks perfecting pitch decks when they should be talking to customers. I've seen great ideas die because the deck wasn't 'investor ready.' This tool gets founders 80% there in 10 minutes so they can focus on what matters.

The Solution

Built an AI powered pitch deck generator with pre built templates, guided sections, and automatic content suggestions based on startup details.

Technical Challenges

AI generated content was too generic sounded like every other pitch. Added industry specific prompts and examples to make output unique. PDF generation with custom fonts and layouts required careful CSS to PDF conversion testing.

Founder Learning

Document generation is harder than it looks. Fonts break, layouts shift, images don't export. Test with real content, not lorem ipsum. And always give users an escape hatch let them edit before exporting.

Results

Functional tool with AI content generation, template system, and PDF export. Shows ability to build document generation apps with structured workflows.

Next.js
TypeScript
Pitch
Generator
Document Generation
Built in 7 days

Cortexreach

21 days

An outreach automation platform with personalization and email automation capabilities.

Why I Built This

Cold outreach is brutal 99% ignore you. Founders waste hours writing personalized emails that get deleted. I built this to prove AI can write emails that sound human and actually get responses, at scale.

The Solution

Created an AI powered email outreach tool that generates personalized messages, manages sequences, and tracks engagement metrics.

Technical Challenges

Email deliverability is complex spam filters killed initial sends. Implemented proper SPF/DKIM setup, warm up sequences, and content scoring. Gmail API rate limits required queue system with retry logic and exponential backoff.

Founder Learning

Email tools live or die on deliverability. You can have perfect AI copy, but if it hits spam, you've built nothing. Infrastructure matters as much as features authentication, rate limits, bounce handling. Not sexy, but critical.

Results

Production ready SaaS with AI personalization, email sequence management, and analytics dashboard. Demonstrates automation workflows and data tracking.

Next.js
TypeScript
AI
Email
Automation
Built in 21 days