After building The Bouncer, I wasn't done with daily games.
Wordle. Connections. Poople. Foximax. Cutle. Zip. Wend. Pinpoint. Crossclimb.
I've somehow become addicted to these tiny games that take up 2–5 minutes of my day. So I wanted to build another one.
But I had a problem. I didn't have an idea. 😭
So I Asked AI
Around this time, I got access to Gemini Pro after buying a Jio data plan. And naturally, my first thought was:
"Well... I've got access to another AI. What can I make it do?" 😭
I decided to use it for something I actually enjoy — brainstorming games. I gave it references from the games I already played and asked it to come up with some ideas.
This was the prompt:
You are a professional mini word-based game developer with a high understanding of how humans reason and what they find exciting. I want to brainstorm some new game ideas similar to Wordle, Connections, Poople, Foximax, Cutle and even LinkedIn's Zip, Wend, Pinpoint, Crossclimb and JeffGoldblumle. The game doesn't have to be extremely unique — it can have similarities with other games. Could you brainstorm a few ideas? Keep the explanation brief, we can go in depth if needed.
And honestly... Gemini cooked. 🔥
It gave me five different game ideas, and I actually liked all five. So I asked it to prioritize them based on concept strength and ease of implementation.
And that's when it gave me: Linkage.
Okay... What's Linkage? 🔗
The concept is simple. You get two words — a starting point and a destination. Your job is to build a chain connecting them using four words from a word bank.
For example:
CENT → Dollar → Green → Colour → Paint → WALLS
The first and last words are given. You figure out the four in between.
There's a catch — you can't type whatever comes to mind. You choose from a word bank, hit Check, and it tells you how many words are correct. Not which ones. Just how many.
No lives. Check as many times as you want.
But there's one thing working against you: ⏱️ Time.
Why I Chose Linkage
Out of the five ideas, Linkage was the one I could immediately visualise.
I wanted that feeling where you look at two words and think "How the hell am I supposed to connect these?" — and then you start experimenting, building a chain, until the whole thing suddenly clicks.
"Ohhhhh."
It was also a really clean concept. No complicated rules, no lives, no unnecessary mechanics. Just:
Start → Think → Connect → Finish.
And admittedly... I also really liked the name. 😂
This Time, I Did Things Differently 🧠
When I built The Bouncer, I made one important mistake — I started with the UI too early.
I wanted to see the game come alive, so I built the interface, made it look pretty, got the animations working... before I had completely figured out the actual puzzle.
That created a lot of ambiguity later.
With Linkage, I did the opposite. Before worrying about buttons or animations, I focused entirely on the puzzle engine and the puzzles themselves. By the time I started on the UI, I knew exactly what I was building around.
The difference was obvious:
The Bouncer's UI is more polished and visually impressive. But Linkage is a better game to play. And that's what actually matters — people aren't opening a daily game to admire the CSS. 😭
The AI Workflow Also Changed 🤖
| Tool | Purpose |
|---|---|
| 🧠 Gemini | Brainstorming and game ideation |
| 🏗️ Gemini | Planning the puzzle engine |
| 💻 Claude Code | Implementation |
| 🎨 Claude Design | UI components and visual refinement |
| ⚡ Superpowers | System design, architecture and planning |
| 📋 Plannotator | Reviewing and approving plans |
Plannotator in particular was pretty cool — it let me review Claude's plan, leave feedback, and approve it before moving forward. That extra layer removed a lot of the ambiguity I had with The Bouncer.
And Then... I Built It In 3 Days 🚀
The Bouncer took around three weeks. Linkage? Three days.
Not because I suddenly became a better coder. The biggest difference was the foundation — I already knew exactly what the game was, how the puzzle worked, what the player would do, and what the core engine needed to accomplish. Very little ambiguity remained.
Once the foundation was solid, building the game became surprisingly straightforward.
Keeping It Simple
Linkage doesn't even need a backend. Puzzles are generated and reviewed locally, then committed as JSON with the frontend. No server, no database request, no extra latency.
Sometimes the best architecture is the one you don't need.
Stack:
- ⚛️ React + Vite
- 🎨 Tailwind CSS
- 🌐 Netlify
- 🐍 Python scripts for puzzle generation
- 📚 DataMuse / ConceptNet where useful
- 🤖 An unhealthy amount of AI
So... Linkage Is Almost Here 🔗
The UI is done. The puzzle engine is working. The word bank is working. The game loop is working. The daily structure is ready.
I just have a few final puzzles to schedule and some last-minute polishing before properly launching it.
And yes — I gave it an admin dashboard inspired by my beautiful admin section™ from The Bouncer. 😌 Apparently I can't build a daily game without one anymore.
Go Try It 👀
If you see the "No Puzzle Today" state right now — no, the website isn't broken. 😭 It's just waiting for me.
Once it's live: two words, four words to connect them, one running timer.
How fast can you build the linkage?
If you play it, I'd genuinely love to know what you think — too easy, too hard, confusing, fun, boring, or if you think I've completely lost the plot. And send it to your friends. See who builds the fastest chain. 🔗
Final Thoughts ❤️
The biggest thing I took away from Linkage wasn't the game itself.
It was realizing how much a solid foundation changes everything.
The Bouncer taught me that the game needs to come first. Linkage gave me a chance to actually apply that lesson. Three days later, I had a playable game.
I started with a random Gemini prompt, five game ideas, and somehow ended up here.
Now I just need to stop getting ideas for daily games before I end up with 47 of them. 😭
See you on the other side of the chain. 🔗
Top comments (1)
Some comments may only be visible to logged-in visitors. Sign in to view all comments.