DEV Community

Jonas Olofsson
Jonas Olofsson

Posted on

I Quit My Job to Build the Game Dev Platform I Wish Existed

#gamedev #vibecoding #founder #bootstrap

I Resigned to Build the Game Dev Platform I Couldn’t Find

About a year ago, I was genuinely excited when I found the first “vibe coding” platforms for games.

The idea felt obvious in the best way.

Describe a game. Let AI handle the rest. Focus on creativity instead of boilerplate.

For a while, it felt like the future had arrived.

By this summer, that excitement had turned into frustration.

How could it still be this bad?

At work, I had already built better coding agents than what these platforms were shipping.

In a hackathon, I built a better animation solution — one that actually worked — and won with it.

Yet the tools claiming to be the future of game creation were still producing:

  • Unplayable prototypes
  • Weak art direction
  • Broken mechanics
  • Systems that collapsed the moment you tried to iterate

They tried to do code and art — but nothing was close to good enough.

And that’s what bothered me most.

It was obvious the underlying AI models were finally good enough.

The limitation wasn’t technology anymore.

It was focus.

I resigned on the spot to build the solution I wished existed.

Why Every “Vibe Coding” Game Platform Failed for Me

Most platforms I tried failed in the same way.

Not because they lacked ambition — but because they optimized for the wrong things.

They were built by:

  • People thinking too much about web design
  • Teams “playing company” too early
  • Products optimized for dashboards and landing pages
  • Demos that looked impressive but weren’t fun to play

They weren’t 100% focused on the agents, systems, and tools that actually create value for game developers.

Games aren’t assets.

Games aren’t snippets.

Games are playable loops.

If movement, enemies, feedback, and progression don’t work together, you don’t have a game. You have a folder.

What I Wanted Instead

I didn’t want another “AI-powered” platform.

I wanted something brutally focused.

  • Go all in on the coding agent
  • Go all in on art tools that are actually game-ready
  • Build a platform where you can create fun, great-looking games

Code and art are just the starting point.

Once quality is there, the real problems become interesting:

  • Music and sound
  • Procedural generation
  • Proper tools for creating content inside the game
  • Systems that make iteration fast and enjoyable

Everything else?

Do the bare minimum.

That philosophy became DreamForge.ai.

What Is DreamForge.ai?

DreamForge.ai is an AI-powered game creation platform designed to turn ideas into playable 2D games as fast as possible.

You describe a game in natural language, and DreamForge generates:

  • Game logic and mechanics
  • Characters, enemies, and environments
  • Animations and VFX
  • A complete gameplay loop
  • A playable build you can test instantly in the browser

The goal isn’t to replace developers.

The goal is to remove everything that prevents people from getting to the fun part: playing, tweaking, and iterating.

Two Small Experiments That Validated the Direction

Once DreamForge started feeling solid, I ran two small experiments.

Not benchmarks.

Not marketing stunts.

Just sanity checks.

Experiment 1: Vampire Survivors–Style Game in 5 Prompts

I wanted to see how fast I could reach a familiar, working loop.

In five prompts, DreamForge generated:

  • A top-down arena
  • Auto-attacking survival mechanics
  • Continuous enemy waves
  • Scaling difficulty

Playable here:

https://dreamforge.ai/play/0a25d9-create-a-complete-vampire-survivors-style-arena-

What mattered wasn’t novelty — it was feel.

The pacing, density, and loop made sense immediately.

Experiment 2: A 2D RPG Boss Fight in 5 Prompts

Boss fights expose everything:

  • Animation clarity
  • Telegraphing
  • Timing
  • Feedback

Again, five prompts.

The result:

  • A focused boss arena
  • Readable attack patterns
  • A fight that’s challenging but fair

Playable here:

https://dreamforge.ai/play/0a25d9-lets-build-a-super-cool-boss-fighting-game-inspi

This was the moment I knew DreamForge was on the right path — not because it was perfect, but because it was playable enough to iterate on.

Why I’m Betting Everything on This

DreamForge.ai exists because I couldn’t find a platform that respected how games are actually made.

Not as code exercises.

Not as asset generators.

But as interactive systems that only make sense once you can play them.

If AI is going to matter in game development, it has to meet creators where the real friction is:

  • Starting
  • Iterating
  • Testing ideas fast

That’s what I’m building.

If You’re Curious, Try It

The best way to understand DreamForge isn’t reading about it.

It’s playing.

If you’ve ever felt like game creation tools were almost right — but never quite there — you’ll probably understand why I resigned to build this.

Top comments (0)