I Built 12 AI Apps in 12 Months: What a Year of Shipping Taught Me

In January 2025, I set myself a challenge that felt equal parts exciting and ridiculous: Build one AI-powered application every month for an entire year.

Screenshot of the Final Review video moments after relase

No breaks. No skipping months. No “I’ll finish it later.”

Just consistent building, learning, and shipping.

Twelve months later, the challenge is complete — and in my Final Review video (embedded below), I walked through the polished versions of all twelve apps, including the final December project that ties everything together. This article is the written version of that journey: the ideas, the builds, the mistakes, the fixes, the lessons, and the reality of committing to a year-long creative and technical challenge.

[embed]

Why I Started the AI Apps Challenge

I wanted to push myself beyond tutorials and coursework. I wanted to build real tools — things that solved problems, things I could show employers, things that would force me to learn new technologies under real constraints. I had already done personal programming projects before, however this was taking them to the next level.

Most importantly, I wanted to prove to myself that I could finish things.

Twelve apps. Twelve months. One developer.

The Journey Through 11 Apps (Before December)

In the Final Review video, I showcased the final versions of the first eleven apps — the polished, fixed, cleaned-up versions. Each one had issues when I first demoed them, but by the end of the year, everything was tightened up.

Below is a short summary of each app — the rest is shown in the video.

January — AI Document Search

A simple but powerful search tool for documents. Cleaned up the UI and improved the upload page styling.

February — AI Testing Agent

Now includes shortened summaries and a direct link to test results.

March — AI Work Hours Calculator

Improved formatting so the AI output is clearer and easier to read.

April — AI Document Extractor & Converter

One of the strongest apps from the start — just refined extraction logic and export formatting.

May — AI Cover Letter Writer

The biggest upgrade of the year: Better formatting, refined tone, CV integration, PDF downloads, improved progress bar, and support for uploading CVs.

June — AI Job Ad Generator

Improved colour contrast and UI consistency.

July — AI Speech-to-Text App

Voice commands now work properly — closer to a real assistant.

August — AI Calendar & Scheduling System

Better email parsing and smoother calendar syncing.

September — AI Document Summariser

Added customisation options and improved summary quality.

October — AI Directory Summariser

Supports template matching, file-type breakdowns, and full directory analytics.

November — AI Programming Assistant

More stable editor, better syntax highlighting, and more accurate AI suggestions.

Each of these apps taught me something different — about UI, about backend structure, about AI prompting, about debugging, and about the importance of iteration.

December — The Capstone: AI Apps Challenge Chatbot

The final project had one goal: Bring everything together.

The December app is a chatbot interface that can:

• Answer general questions
• Switch between prompt modes (code, writing, Q&A, etc.)
• Call functions from any of the previous 11 apps
• Open the previous apps in 3 modes: In-chat, Quick Form modal and Full App (Opens the full app in a new tab)

Ask it to summarise a document? It calls the September app. Ask it to generate a cover letter? It calls the May app. Ask it to analyse a folder? It calls the October app.

It’s the “hub” for the entire challenge — a single conversational interface that unifies a year’s worth of work.

To make this possible, I restructured the project (Check out the GitHub repo) so the chatbot lives in the root directory, and all helper functions from the previous apps can be imported cleanly. The 12.December folder contains the chatbot logic, routing system, and prompt mode handlers.

This project felt like the final boss — the moment where everything I’d built suddenly connected.

Deployment Plans

I haven’t deployed all twelve apps yet — but that’s the next step.

The plan is to:

• Containerise each app with Docker
• Deploy them to AWS, Azure or Google Cloud Platform
• Create a clean portfolio site linking to each one
• Add documentation and usage guides

Deployment is its own challenge, especially with twelve apps that all use different models, libraries, and file-handling logic. But it’s coming — and it will make the entire challenge accessible to anyone who wants to try the apps themselves.

Challenges I Faced

This challenge wasn’t smooth. At all.

Some of the biggest hurdles:

• Time pressure — building an app every month while studying and working
• API changes — especially the Gen AI SDK deprecations
• UI consistency — early apps looked rough compared to later ones
• Debugging — especially file parsing, audio processing, and calendar syncing
• Feature creep — wanting to add “just one more thing” every time

But the hardest part? Showing unfinished work publicly. Every video included bugs, missing features, or rough edges — and that was uncomfortable. But it forced me to iterate, improve, and finish.

Lessons Learned

Here’s what a year of building taught me:

1. Consistency beats perfection

Shipping something imperfect is better than never shipping at all.

2. Structure matters

By the time I reached the December project, I realised how much easier everything becomes when earlier work is modular and organised.

3. AI development is experimentation

You try things. They break. You adjust. You try again.

4. Creativity grows with constraints

A monthly deadline forces you to make decisions, cut features, and focus on what matters.

5. You learn more by building than by reading

Every bug, every UI issue, every broken API call taught me more than any tutorial could.

Closing Thoughts

The AI Apps Challenge has been one of the most rewarding projects I’ve ever done. Twelve apps. Twelve months. Countless lessons.

If you want to see the full journey — including the polished versions of all apps — check out the Final Review video (linked near the top of this). It’s the best summary of everything I built this year.

And if you’re thinking about starting your own challenge — do it. You’ll learn more than you expect, and you’ll finish with something you’re genuinely proud of.

Here’s to whatever comes next.

Relevant Links

Python Tutorials Website

Link to Portfolio Page

Python Youtube Channel

Python Programmer Instagram