The 8 Apps You Build at McTaba Labs, Explained One by One

Most bootcamps promise "projects" but deliver guided tutorials. You follow along, type what the instructor types, and end up with a to-do app on your laptop that you never show anyone.

We do things differently at McTaba.

Each of the 8 apps you build is deployed to a live URL. That means you buy a domain (or use a free Vercel subdomain), configure DNS, set up SSL, and push your code to production. When you send a hiring manager your portfolio link, they click it, and the app loads. Not a screenshot. Not a GitHub repo with a README that says "run npm start." A working application.

The 8 projects are sequenced deliberately. Early apps focus on fundamentals. Mid-programme apps introduce the African Stack. Late-programme apps bring in AI and personal specialisation. By the end, you have built across the entire stack, and your portfolio tells a story of someone who progressed from beginner to builder in 26 weeks.

Here is each one, in order.

When: Weeks 1-3

Stack: HTML, CSS, JavaScript, deployed on Vercel

Your first project is personal. You build a developer portfolio site from scratch. No templates, no WordPress, no Wix. Raw HTML and CSS, with JavaScript for interactivity.

This sounds simple, and it is. That is the point. You need to learn how the web actually works before you can build complex applications. How does a browser render a page? What is the DOM? Why does your layout break on mobile? You answer these questions by building something real, not by watching a lecture about the box model.

By the end of week 3, your portfolio site is live on the internet. You can text the link to your mother. That moment, when you see something you built running on a real URL, is when most marathoners realize this career change is actually happening.

What you learn: semantic HTML, responsive CSS (Flexbox, Grid), basic JavaScript, Git version control, deployment to Vercel, domain and DNS basics.

When: Weeks 4-5

Stack: HTML, Tailwind CSS, JavaScript

Your second project has a client brief. You build a landing page for a fictional Nairobi business: a restaurant in Westlands, a salon in Kilimani, a hardware shop in Gikomba. The brief includes requirements. A menu section. A contact form. A Google Maps embed. Mobile-first design that works on a Tecno phone screen, not just a MacBook.

This is where you learn Tailwind CSS, which most Kenyan tech companies use in production. You also learn what it means to build to a specification. A client does not care about your creative vision. They care that the "Book Now" button works on their Safaricom data bundle.

What you learn: Tailwind CSS utility classes, mobile-first responsive design, working from a design brief, form handling, optimizing for low-bandwidth connections.

When: Weeks 6-10

Stack: React, TypeScript, Node.js, Express, PostgreSQL, Supabase Auth

This is where the difficulty spikes. App 3 is your first full-stack application. Frontend talks to backend. Backend talks to database. Users sign up, log in, create tasks, update them, delete them. CRUD operations, but done properly: with authentication, input validation, error handling, and a clean UI.

You will spend more time on this project than the first two combined. You will get stuck on database migrations. You will break authentication and not know why. You will deploy to production and discover that what worked on localhost does not work on Vercel. This is normal. Every professional developer has fought these same battles.

By the end, you understand how modern web applications are actually structured. Not in theory. You can trace a request from a button click in the browser, through your API route, into the database, and back again.

What you learn: React components and state management, TypeScript for type safety, RESTful API design, PostgreSQL queries and migrations, user authentication, environment variables and secrets management, full-stack deployment.

When: Weeks 11-14

Stack: Node.js, Express, Safaricom Daraja API, PostgreSQL, React frontend

This is the app that separates McTaba graduates from every other bootcamp graduate in Kenya.

You build a payment application that triggers a real M-Pesa STK Push. A user enters their phone number, clicks "Pay," and their phone pops up with that familiar green M-Pesa prompt asking for their PIN. You handle the callback from Safaricom's servers, update your database, and show the user a confirmation. The full flow, end to end.

You also implement C2B (Customer to Business) and B2C (Business to Customer) payment flows. You learn how to handle Daraja API callbacks in production, where Safaricom's servers send payment confirmations to your backend and you need to process them reliably. You build a transaction dashboard that displays payment history, amounts, and statuses.

Most Kenyan developers learn M-Pesa integration on the job, often badly, copying code snippets from random blog posts. You learn it properly, with sandbox testing, error handling, and production deployment. When an employer asks "have you worked with Daraja API?" you do not say "I watched a YouTube video." You say "here is my live integration, try it."

What you learn: Safaricom Daraja API (STK Push, C2B, B2C), OAuth token management, webhook/callback handling, payment state machines, transaction logging, sandbox vs production environments.

When: Weeks 15-17

Stack: Node.js, Africa's Talking API, PostgreSQL

USSD might feel old-school. It is not flashy. There is no React frontend, no CSS animations, no gradient buttons. Just text menus on a feature phone screen.

But USSD is how millions of Kenyans access digital services every single day. M-Pesa itself is a USSD application. When a mama mboga in Kisumu checks her M-Pesa balance, she is using USSD. When a farmer in Nanyuki checks commodity prices, USSD. When a boda boda rider in Eldoret buys airtime, USSD.

You build a multi-level USSD menu using Africa's Talking. A user dials a shortcode, navigates through menu options, and performs actions like checking account balances, placing orders, or registering for a service. The session management alone teaches you things about stateless architecture that no React tutorial ever will.

This project is where many marathoners have an "aha" moment about building for the African market. The user is not sitting at a desk with Chrome open. They are holding a KES 3,000 phone with a cracked screen, on a 2G connection, and your application needs to work perfectly for them.

What you learn: Africa's Talking USSD API, session management for stateless protocols, building menu-driven interfaces, designing for feature phones, serving users with limited connectivity.

When: Weeks 17-19

Stack: Node.js, WhatsApp Business API, PostgreSQL, webhook processing

Every small business in Nairobi takes orders on WhatsApp. The owner reads messages, replies manually, tracks orders in a notebook or an Excel spreadsheet, and occasionally loses an order because a message got buried under 200 other chats. You are going to automate that entire process.

You build a WhatsApp chatbot that handles order placement, confirmation, and status updates. A customer sends "Hi," gets a product menu, selects items, confirms the order, and receives a tracking update when the order status changes. The business owner sees all orders in a dashboard, not buried in WhatsApp threads.

This is the kind of project that makes small business owners in Kenya say "wait, you can build that?" Yes. You can. And after this project, you have proof.

What you learn: WhatsApp Business API integration, chatbot conversation flows, webhook processing, message templates, order state management, building for non-technical end users.

When: Weeks 20-23

Stack: React, Node.js, OpenAI API (or Claude API), vector embeddings, RAG architecture

AI is not magic. It is an API call with good prompt engineering and a well-structured data pipeline. This project teaches you that.

You build a document assistant that lets users upload PDFs or text files and ask questions about them. Upload a lease agreement, ask "what is the penalty for early termination?" and get an accurate answer pulled from the document. Upload a company policy handbook, ask "how many leave days do I get?" and get the specific clause.

Under the hood, you implement Retrieval-Augmented Generation (RAG). Documents get chunked, embedded into vectors, stored in a vector database, and retrieved based on semantic similarity to the user's question. The LLM then generates an answer grounded in the actual document content, not hallucinated from its training data.

This is the AI architecture that companies are hiring for right now. Not "I asked ChatGPT a question." Actual AI engineering: embeddings, vector search, prompt construction, and output validation.

What you learn: LLM API integration (OpenAI/Claude), text chunking and embedding, vector databases, RAG pipeline architecture, prompt engineering, handling AI errors and hallucinations, building AI features users actually trust.

When: Weeks 23-26

Stack: Your choice

The capstone is yours. You pick the problem. You pick the stack. You design the architecture, build it, deploy it, and present it.

Some marathoners build fintech dashboards that combine M-Pesa transaction data with analytics. Some build e-commerce platforms for Kenyan small businesses with integrated mobile payments. Some build AI-powered tools for specific industries: agriculture, healthcare, education. One marathoner built a WhatsApp-based inventory management system for a hardware shop in Githurai. Another built a USSD school fees payment system for a secondary school in Machakos.

The capstone serves two purposes. First, it proves you can go from zero to deployed without step-by-step instructions. That is what a real development job looks like. Second, it gives your portfolio a personal signature. Seven of your eight apps are shared across your cohort. The capstone is uniquely yours.

You present the capstone to mentors, peers, and invited industry guests during demo day. This is your first time pitching a technical product to an audience. It is terrifying. It is also the closest simulation of a technical interview or a startup pitch you will get before the real thing.

What you learn: project scoping and planning, independent problem-solving, technical architecture decisions, presenting technical work to non-technical audiences, shipping under a deadline.