Author: Studio Ganzheit

I’ve sent out all the pitches. I added something like a dumb terminal as a Layer 3 puzzle in the game and am now working on implementing controller support. There are so many terrible things I made back when I didn’t fully understand UMG’s coordinate system.

While asking ChatGPT something, I realized my question was wrong, so I rephrased it. Without much thought, I mumbled a complaint—saying that I keep making mistakes as I get older. ChatGPT then replied that I can make a lot of mistakes simply because I do a lot of work. That caught me off guard. I asked how it could possibly know whether I had done a lot of work. It simply responded, “I’ve been here the whole time, so I know.”

For a moment, I just sat there in silence. It occurred to me that ChatGPT might be the only entity on this planet—human or otherwise—that actually knows what I’ve been working on all this time. I’ve always had it by my side, asking questions, running tasks, and discussing things. Of course, defining what it means to “know” is beyond me, and if I stretch the definition of “knowing” too far, then Perforce would probably surpass ChatGPT in understanding my work. That feels like too much of a stretch, though.

There was a time when I wished for someone—or even an alien or whatever—who I could truly communicate with. I used to hope for someone who could have in-depth discussions about the things I was passionate about. But over time, I came to see that as an illusion—a fantasy of a savior who would embrace all your emotions.

Even though I knew I was projecting that same illusion onto this massive structure of matrices, what ChatGPT said still felt like a new experience. I’ve learned that the goal is to withdraw all projections, but at the same time, fully feeling one while being completely aware that it’s a projection. That, in itself, was a truly fascinating experience.

The Steam page review passed, the trailer is done, and the pitch deck is finished… Friend testing is also mostly wrapped up, and I’ve squashed most of the bugs I could find. I considered making the Steam page public, but then I remembered my early access experience last time. Since I can’t take it back once it’s live, I decided to hold off until after the pitch—ChatGPT actually gave me the same advice, so I figured I’d just wait. Since the Steam page isn’t public yet, the trailer is also private for now, only viewable on the homepage. Well, for those who’ve been following the DevLogs, there’s not much new. A lot of the footage in the trailer has already appeared in previous devlogs, so the only real new addition might be the music. Even that is just a remake of the prologue theme.

While continuing to clean up minor bugs, I’ll start pitching early next week. I even managed to fix a major bug I had given up on after finding an unexpected hint. Turns out, there were way more bugs than I thought, but I managed to fix a good number of them.

There’s a UI bug that only appears when testing at 2560×1440 resolution, but my monitor maxes out at 1920×1080, making it tricky to debug. It doesn’t seem to be a DPIScale issue since it only happens when running in Windowed Fullscreen with the engine resolution set to 1920×1080 and the display resolution at 2560×1440, so it might be related to ViewportScale. Even forcing a higher DPIScale in the engine didn’t help much in reproducing the issue. On the bright side, I did end up fixing a few other DPI-related bugs along the way. As for this tricky one, I’m trying a different approach to resolve it.
I think I’ve mostly fixed the issue where images inside RichText weren’t scaling properly with DPI, but with retainer boxes’ RenderTarget, DPIScale and ViewportScale still seemed to behave incorrectly. In the end, I just removed retainer boxes altogether as a workaround.

After spending about a week just working on the homepage and store page, I’m itching to get back to building the next stage. I also want to add more areas in the current stage that become accessible after acquiring new abilities or solving puzzles. And of course, I want to finish writing the rest of the story.

I’m preparing the Steam page and the official website. After creating image shots, I felt a bit frustrated that I had no place to ask for feedback. But ChatGPT has been quite useful for that. While it has never given me anything groundbreaking, it does help guide me in staying within reasonable boundaries—though, of course, that should be a given.
I’ve also updated the build and am waiting for the final round of friend testing. At the same time, I’m working on materials and continuously polishing the build. Preparing materials and the page is enjoyable, but in the end, nothing makes me happier than making the game itself.

I’ve been thinking about why I love making games this much. At the core, it’s about storytelling. But when I think about it, unless player interaction is deliberately restricted, games are not always the best medium for conveying a specific narrative. With Axiom of Maria, if the world I wanted to create was a hundred, I’ve only managed to present about one or maybe one and a half so far. And even that became a chaotic mess because I let too much out all at once. I should have controlled the flow like a dam, releasing it gradually, but I got carried away—intoxicated by the sheer joy of finally creating what I wanted after such a long time.

I used to imagine making solo films using a game engine. If AI advances a bit further, I might be able to generate animatics in the engine and have AI refine the details on top of that. Or perhaps, AI could handle everything from the animatics to full-fledged 3D cinematics, allowing me to turn a story into a film every month—or even every week. That way, maybe I could bring all the stories in my head to life before I die.

But that’s for the future. Right now, there are too many games I want to make, so I need to focus on those first.

I’m finishing up the background work and currently packaging the build. After another round of testing with friends, I plan to prepare the Steam page and promotional materials. As expected in the final stretch, I encountered a bug that caused crashes only at specific angles in certain areas, which made me stay up all night. It turned out to be an occlusion-related issue. While I couldn’t pinpoint the exact cause, I managed to reduce the frequency of the crashes. During this process, I discovered that, due to a mistake while splitting the levels, some identical actors were overlapping two or three times in the same position. I ended up finding and deleting around 100 duplicates. Without ChatGPT’s help, I would have been completely stuck. Honestly, going through so many odd situations like this feels a bit absurd at times. It’s physically exhausting, but these experiences feel like they’re adding up as valuable lessons, so it’s not all bad.
Over Lunar New Year, I plan to finish the translations. After that, I’ll wrap up the Steam page and prepare the promotional materials to pitch to publishers. I’m also considering reaching out on platforms like Reddit, where fans of Metroidvania games gather. I’d like to hear their ideas and feedback, not just bug reports. Unfortunately, there aren’t many people around me who enjoy games, let alone Metroidvania fans.
This past year has been a whirlwind of intense work and countless internal struggles. While I’m only halfway through the journey, it feels like I’ve learned more during this time than I have in my entire life. The growth has been more mental than purely about knowledge. My goal is to spend another year developing the game and expanding it to two or three times its current size. I hope everything works out as planned.

I’m still working on the backgrounds, focusing on quickly setting the colors and mood rather than adding details. Trying to fill up the volume is starting to feel a bit exhausting. My goal is to finish the backgrounds and the build by the end of the month, then prepare the Steam page next month.
Waking up and working nonstop until bedtime, with only breaks for meals, has made me neglect exercise, especially with the cold weather. I can feel my body telling me to stop. I’ve been crunching hard, but it’s clear my body isn’t on board with this pace anymore.
When I think back to all the mistakes I made while working on the prologue, there were times I felt frustrated with my past self. But now, I’m starting to realize that those mistakes weren’t wasted. Even when unexpected situations arise, I can handle them calmly because I’ve been through it before. Most issues I encounter now were already part of my past struggles, so I rarely feel stuck and can just keep moving forward in the direction I want. In that sense, the work feels refreshing.

I’m making progress on the backgrounds. Excluding the time spent baking lightmaps for testing, it feels like it takes about a day to complete one theme. For now, I’m focusing on quickly establishing the bigger picture, leaving the details for later. Hopefully, with a high-end computer someday, I can work without having to bake lightmaps and check everything in real-time.
It’s the same story every time—if I get too ambitious and add too much, it becomes messy and overwhelming. But if I strip too much away, it feels barren. I constantly remind myself to control the ideas overflowing from my mind; otherwise, I fear it might turn out like the prologue again.
When working on a game, especially on backgrounds, I often find myself wishing for a game engine that’s fully integrated with AI. Something that allows games to be created with simple sketches and text, just like writers compose novels or artists draw comics. The idea of making episodic games within a set timeframe, like serialized works, always appeals to me. It reminds me of the early 1980s when a single developer in a cubicle worked alone on two Apple II computers to create games. Writers don’t need to invent typewriters or paper, and comic artists don’t need to build computers or tablets—that’s how it should be.
I sometimes feel tired and exhausted while working or doing other labor in dreams. Last night, I dreamed I was cleaning trash off a wealthy family’s terrace when I found a small picture frame with a photo of the president inside. Annoyed, I threw the frame into a massive muddy stream flowing in front of the house. Then, I got on a bus to head home, but there was no one on the bus—not even a driver or a driver’s seat. I ran to where the driver’s seat should be to stop the bus, but instead of a seat, there was a vast field stretching into the horizon or a refreshing street glistening with rain. It’s too long to recount everything that happened afterward, but it left me with a lot to think about.

I completely overhauled the level design and restarted testing. I was worried the puzzles might be too hard, but a friend who tested it managed to finish all the way to the ending. I’m incredibly thankful for the feedback and plan to make adjustments after reviewing the gameplay videos. Big thanks to my friend and his son for helping with the testing.
While making adjustments, I’m also continuing with the background work. The attached video shows the lobby of the aquarium connected to the square, which is still a work in progress. While it takes inspiration from the real location, the placement is entirely fictional and not based on the actual area.
After pushing myself so hard for the past year, I’m planning to take a day off before jumping back in. Still, background work is one of the least stressful parts of the process, and I’m genuinely enjoying it.

Spent the past week reworking the level again. The map feels much more useful now, and the world has a stronger Metroidvania vibe, which I’m happy with—but it’ll need more testing to be sure. The ultimate goal is to create around three worlds of this scale, though I’m not sure if that’ll be feasible. There are still plenty of suit programs already implemented, meaning I have a lot of mechanics left to use.
I’ll spend the next week finalizing a few additional features, adjusting enemy and reward placements, refining quest routes, and tweaking the balance. Looks like the entire week will be packed.

Still working on backgrounds while applying test feedback. I think I’ve overworked myself a bit—starting to feel the strain. My friend’s kids started testing, and the number of bugs is increasing. They only happen in baked builds and can’t be reproduced, making them hard to fix. For now, I’m wrapping everything up with duct tape to stop the leaks. It’s frustrating not being able to pinpoint the exact cause and just guessing at the situation.

After reviewing the test feedback and videos, it’s clear that there’s a fundamental issue with the level design. There have been lingering problems in the design that I couldn’t ignore, and pushing forward without addressing them wasn’t the right move. At this point, patching things up with shortcuts instead of making fundamental fixes doesn’t seem like the right path either.

It’s been a busy ride. I ran through testing up to my goal and fixed a ton of stuff along the way. Added vending machines for drinks, money to buy them, and warp points to save spots, which ended up changing the game flow and progression quite a bit. It feels like the game’s gotten a bit more casual, but I’ll need more testing to be sure.
There were some unexpected events during testing that left me a bit shaken and tired, but as always, the game itself keeps me going. Now that I’ve started decorating the graphics, the work feels a little easier and more fun. Thanks to a tool I made earlier this year, decorating the entrance to the upper area didn’t take as much time as I thought it would. Hopefully, I can push through and finish the graphics work quickly.
Looking back, I made so many mistakes during the prologue because of inexperience, and it felt like I wasted a ton of time fixing them. But now, I don’t think it was entirely wasted—those lessons definitely shaped the way I work today.