Project #1 MR Recommender, Life is too short for a bad movie 🎬

Link to the Product: MR Recommender

What movie are you watching tonight?

I usually end up texting my brother for recs. A few weekends ago, though, while waiting for Luigi to reply, I thought: why not try building something with AI? I wanted to start learning no-code tools and this was the perfect small problem I could try to solve.

🎥 Enter MR Recommender: a smart AI movie recommender built in a weekend (with no-code tools)

👉 mrrecommender.lovable.app

We’ve all been there: 30 minutes into scrolling Netflix and still no clue what to watch. So I finally gave no-code a shot. I got the first version of the recommendation engine live even before Luigi texted back. Answer a few questions, and you will be recommended up to 5 movies personalized on your preferences.

Tools I have used to build the tool

• 🧡 Lovable – to build the frontend

• 🧠 OpenAI – to power the recommendation logic

• 🎥 OMDb API – the IMDb of APIs

• 🗂️ Supabase – for managing API keys

💡 What I learned building it (and why you should also give no-code a shot):

1) No-code is real: Lovable is ridiculously fast to get started with—way easier than Shopify, Webflow, or Squarespace for simple ideas.

2) Prompting is the product: The speed is amazing, but can also be frustrating, especially on the design/UX side. I initially asked for a Masterclass-style layout… then ended up changing pretty much everything (including colours) because I didn’t love it.

3) The last 20% is 80% of the work: I spent 80% refining the tool. While Lovable gives you the recommendation engine fast, most of the work is really understanding how it was built and tweaking it.

4) Prototype ≠ final product: To build a truly smart recommendation engine, I do think you need more sophisticated technical understanding. But honestly, I loved the process of prompting Lovable and tweaking the logic to get something working quickly and learning. 

Why this matters:

As a non-technical person, the gap between ideas and products often felt too wide. But now? That gap is closing ridiculously fast.

We talk a lot about AI and no-code, but nothing beats actually building something. Even if it’s small and a little rough like Mr. Recommender, you’ll learn a ton. I’ve always been a big believer in learning by doing and in the motto  "Done is better than perfect”