I built a free bitmap font generator that converts any reference image into Unity/Godot-ready font sheets
Hey IH! I'm a Spine animator at a game studio — not a developer — and I just launched a small web tool that solves a real workflow pain.
Our designer kept generating beautiful font styles with AI image tools, but none of the outputs could be directly imported into Unity or Godot. Game engines need a proper BMFont sheet (.png + .fnt file) to use bitmap fonts — and converting AI output into that format was a huge manual headache.
So I built CopyPxl to bridge that gap:
Upload any reference image → get a complete bitmap font sheet in BMFont format, ready to import into your game engine.
What it outputs:
- font.png (sprite sheet with all characters A-Z, 0-9, symbols)
- font.fnt (BMFont descriptor file)
- ZIP download
Works with: Unity (TextMeshPro + Legacy GUI), Godot, Phaser, LibGDX
Free to start — sign in with Google and get 10 credits.
For context on how I built this: I'm not a web dev. I started with a messy Python script, then slowly turned it into a full web app with a lot of iteration (and Claude's help for the coding parts). The whole journey from "frustrated at work" to "live SaaS product" took about a month.
Happy to answer any questions — especially from game devs who've dealt with custom font pipelines!
Joongix, bridging the gap between AI-generated art and engine-ready assets like .fnt files is a massive win for game dev workflows. By automating the conversion of reference images into structured BMFont sheets, you've effectively removed the "manual headache" that usually kills the momentum of creative experimentation in Unity and Godot.
I’m currently running Tokyo Lore, a project that highlights high-utility logic and validation-focused tools like CopyPxl. Since you’re building the definitive bridge between static design and interactive implementation, entering your project could be the perfect way to turn this "frustrated at work" origin story into a winning case study while your odds are at their absolute peak.
Thanks for the kind words, Tokyolore! You nailed the core problem — it's not just about generating art, it's about getting it into a format the engine can actually use. That last-mile conversion is where most tools stop.
Tokyo Lore sounds interesting — I'll check it out. Always good to see projects focused on helping builders find the right tools.
Appreciate it — and yeah, that “last-mile” gap is where most things quietly break.
Curious — are you focusing more on a specific engine/workflow right now, or keeping it flexible across use cases?
Also, I’m running a small project (Tokyo Lore) where we highlight tools solving exactly these “last-mile” problems and get them in front of a focused group of builders.
Since your angle is pretty sharp, it could stand out well in this round — especially while it’s still early.
It’s a $19 entry, includes a structured analysis + entry into a live round (winner gets a Tokyo trip — flights + hotel).
Happy to share more if you’re interested 👍
Keeping it flexible by design — the output is standard BMFont format (font.png + font.fnt), so it drops into anything that supports BMFont: Unity TextMeshPro, Godot's BitmapFont importer, Phaser, LibGDX.
In practice Unity and Godot are the most common use cases coming in, so those get the most testing attention. But the format is intentionally engine-agnostic.
The "last-mile" framing is exactly right — that's the gap CopyPxl fills. Tokyo Lore sounds interesting, I'll take a look.
That makes a lot of sense — BMFont as the layer keeps it clean and future-proof 👍
And yeah, Unity + Godot being the main pull makes sense — probably where the pain is sharpest too.
The “last-mile” angle is honestly your strongest hook. Most tools stop at generation, you actually finish the job — that’s what people pay for.
No rush on Tokyo Lore — take a look when you get a chance. If you decide to join later, your use case would fit really well, especially with real workflows behind it.
Really appreciate the kind words, Tokyolore! And good callout on the framing — that "last-mile" problem is exactly what drove me to build this.
One quick update on my end: I had mentioned the Google login issue was resolved, but I spoke too soon. There was still a credit assignment bug that's now fully confirmed fixed as of today (May 3). Anyone who signed in with Google and didn't receive their 10 credits — please try again, it should work now.
And yes, will definitely check out Tokyo Lore when I get a chance!