To learn more about how Flyknight came to life, Axel García, Head of Brave Zebra, sat down with Miguel for a candid conversation about solo development, technical challenges, community building, and what it really takes to make an indie game that connects with players.
Chapters of the Interview:
- 00:00 — Intro & how Flyknight started as a Game Jam project
- 02:19 — From outsourcing to indie: Miguel’s background and skillset
- 03:44 — The challenge of building systems in Blueprints
- 04:49 — Working with a small team and learning to delegate
- 06:45 — 200,000 copies sold: unexpected commercial success
- 09:03 — How Twitter helped build momentum before launch
- 12:31 — Marketing without cringing: posting in a way that feels natural
- 13:18 — Lessons for other indies: letting go of perfection and working smart
- 19:56 — Why a world of humanoid insects made sense—and worked
- 25:41 — Visual style: unlit rendering and skipping polish to save time
- 27:10 — Designing enemy AI with State Machines instead of Behavior Trees
- 30:22 — Co-op chaos: how multiplayer changed the feel of the game
- 32:25 — Accessibility, short playtime, and price point as success factors
- 43:10 — What’s next for Flyknight: randomizer mode, console ports, sequel
- 45:41 — Returning to Sorceress after Flyknight’s success
- 52:37 — Persistent technical limitations and the cost of solo dev
- 1:00:41 — Final thoughts: make games you believe in, not just what sells
All the questions
Could you briefly introduce Flyknight?
Flyknight is a first-person action game inspired by the older and more obscure titles from From Software, like King’s Field and Shadow Tower Abyss. It actually started as a submission for a Game Jam with the theme “bugs,” which is why the characters in the game are humanoid insects. At the time, I was feeling a bit burned out from working on my main project, Sorceress, which I had been developing for several years. So I used the jam as a way to disconnect and try something new—something smaller, faster, and weirder.
The idea was to make a short, focused experience with a retro vibe, simple mechanics, and some fun combat. But even during the jam, it started to get a surprising amount of attention online. That gave me the push to turn it into a complete game. What started as a 40-minute prototype eventually became a 3-hour co-op action RPG that I finished in around six months.
What’s your background before going indie?
I studied art at ESAT, a design and tech school in Valencia. After that, I worked a few jobs here and there—one was doing visualizations for architects, and another was at a tech company called Aris Nova. But the job that probably had the biggest impact on me was at Ocean Studios, an outsourcing company in Valencia. I was there for three years, from 2019 to 2022. I started as a junior artist and eventually ended up leading a small project.
That experience gave me a good understanding of the production side of things—and also made me realize how much time is often spent in AAA or outsourcing environments on ultra-polished assets, sometimes at the expense of creativity or morale. I think that experience shaped how I approached Flyknight. I didn’t want to spend months perfecting small details. I wanted to build something playable and fun, even if a bit rough around the edges.
You mentioned doing everything in Blueprints. Is that still the case?
Yes, Flyknight was built entirely in Unreal using Blueprints—no traditional coding. I come from an art background and never really learned to program properly, but I’ve managed to get by through trial and error, watching tons of tutorials, and just experimenting.
That said, this project pushed me way outside my comfort zone. Adding multiplayer, inventory systems, saving and loading—it got very technical, very fast. At one point, I even caused people’s save files to overwrite each other because I misunderstood how networking worked. So yeah, Blueprints helped a lot, but it was still a huge learning experience.
Did you develop the whole game by yourself?
Not entirely, though a big chunk of it was on me. I handled all the programming and design, but I worked with three friends who helped shape the project. One did all the 3D art, another handled the music, and the third worked with me on the narrative, while also acting as an unofficial producer. He kept me grounded whenever I started drifting into feature creep—which happens a lot when you’re used to working on long, system-heavy games like Sorceress.
I could’ve done the art myself—I studied it and have industry experience—but I wanted to finish the game quickly. Delegating meant we could move faster and, honestly, the artist has better taste than I do. Letting go of that part also gave me space to focus on what I still needed to learn: making the whole thing function.
So in the end, Flyknight was a four-person collaboration, though the production structure was very lightweight. Everyone contributed in a focused, efficient way, and that’s probably why it worked.
What was the biggest challenge during development?
Honestly, I went into Flyknight thinking it would be a quick, low-stress project. After years of working on Sorceress, which is full of complex systems and interconnected mechanics, I wanted something simpler—“you hit an enemy, it takes damage, maybe loses a limb.” That kind of thing.
But then I decided to add multiplayer, inventory, and save/load functionality, and that’s when things exploded in complexity. I thought it would be manageable because Unreal makes certain multiplayer features fairly accessible. But that assumption was… overly optimistic. Debugging sync issues, figuring out server-client communication, replicating variables—all of that became a massive headache. For example, one of the worst bugs at launch was a crash that only happened when three or more players cast spells at the same time. I had never tested for that scenario because I only ever tested with one other person.
And there were weirder ones too—like players losing items from chests, or accidentally overwriting their save files with someone else’s. Looking back, it’s kind of wild how many things broke, and how forgiving players were during that first wave of patches.
How did the co-op feature come about?
It wasn’t part of the original plan. The first version of the game—what I made for the Game Jam—was single-player only and lasted maybe 30–40 minutes. It didn’t have multiplayer, inventory, spells, parrying, fishing, or anything like that. Just basic combat.
But as I started expanding it, multiplayer became something I was curious about. I had played a lot of Dark Souls and Elden Ring with friends, and I loved the chaos that co-op adds to otherwise serious games. So I decided to give it a shot, thinking it wouldn’t be too hard to implement. Spoiler: it was.
Despite the technical pain, I’m really glad I added it. The difference between playing Flyknight alone versus with friends is huge. In single-player, it’s moody and atmospheric. In co-op, it turns into a comedy show. There’s friendly fire by default, players can dance mid-battle, and the game even changes the music if you start dancing. It’s ridiculous and fun in a very different way, and that contrast adds a lot to the experience.
What do you think made the game resonate with players?
I think it was a mix of factors—none of them individually spectacular, but together they made the game accessible, distinctive, and shareable.
For one, the visual style helped. The game has no lighting, no textures, and everything is vertex-painted. It’s completely unlit, which gives it a clean, low-poly look that feels retro without being cliché. Most retro-style indies either go for PS1-style crunchy textures or pixel art. I went in a different direction—almost like RuneScape with a bit of Shadow Tower Abyss weirdness—and that caught people’s attention.
Also, the price point made a difference. At launch, Flyknight was €5.99. Some people told me I should’ve charged more, but for me, the priority was accessibility. It’s easier to convince a group of friends to try something new if it’s cheap. Plus, the game runs on low-spec machines and offers a full experience in just a few hours. That made it feel very “low-commitment,” which probably helped with word of mouth.
And then there’s the co-op. I really think that’s what pushed the game beyond niche. Playing with friends naturally leads to funny moments, and those moments get shared. That virality can’t be manufactured, but it definitely helped Flyknight travel faster than expected.
Did you expect it to be this successful?
Not at all. When I launched the game, it had about 16,000 wishlists—a decent number for an indie title, but not a sign of a smash hit. I assumed that maybe 20% of those would convert to actual sales, which would’ve already made the whole thing worth it for me.
Instead, the game took off. That first weekend blew past my expectations. It wasn’t just the sales, though those were great—it was the fact that people were genuinely enjoying the game. They were posting about it, making videos, sending me kind messages… it was a bit surreal.
But I think the key was that the game hit the right balance: cheap, fun, weird, and short. That combo made it easy for people to try, and then recommend.
How important was Twitter to that success?
Very. I’ve been on Twitter since around 2017–2018, mostly posting about Sorceress, my long-term project. Over time, I built a small following—nothing huge, maybe 6,000 to 8,000 followers—but they were engaged. When I started posting GIFs and clips of Flyknight, those tweets got 2,000 to 4,000 likes consistently.
Of course, likes don’t equal sales. But that level of attention told me people were curious. It gave me the confidence to keep going and confirmed there was an audience out there for this strange little game.
Also, and maybe more importantly, Twitter helped me find my voice. I don’t enjoy traditional marketing—I find it a bit cringey—but I do like sharing weird or funny clips with friends. So I just leaned into that and treated Twitter like a place to share progress with people who might care. That mindset helped me stay consistent without burning out.
What advice would you give to other indie developers?
I think the most important thing is to find a way of working and communicating that feels natural to you. There’s a lot of advice out there that sounds like hustle culture—post daily, follow trends, optimize everything—but that only works if it doesn’t burn you out. Personally, I find traditional marketing a bit awkward. So instead of forcing myself to “sell,” I just shared things that I found funny, cool, or weird. That made it sustainable.
Another piece of advice: don’t try to do everything yourself. I used to think that being a “true indie dev” meant doing it all—code, art, design, marketing. But that mindset can be counterproductive. You end up overwhelmed, progress slows, and it’s easy to lose motivation. With Flyknight, I gave up control over the art and music, trusted my collaborators, and finished the game faster and better than if I’d done it solo.
So yeah, be realistic. Ask for help. Do what you can with what you have—and know when to let go of control.
From a technical perspective, what are you most proud of?
Probably the fact that the entire game runs in Unlit mode—no lighting, no shadows, no post-processing. It doesn’t even react to directional lights. That decision started as a creative constraint, but it ended up being one of the smartest things I could’ve done. It gave the game a unique style, simplified production, and helped hide limitations.
Another thing I’m happy with is how I handled b>enemy AI. Instead of using Unreal’s standard Behavior Trees, I went with a State Machine system using a plugin. It’s not revolutionary, but it let me design AI behavior in a way that felt more readable and flexible. For example, an enemy might patrol, chase you, attack, get stunned, and recover—each of those is a state, and the transitions between them can be customized easily.
This also allowed me to play with unpredictability. I was inspired by how weird some enemies are in Shadow Tower Abyss—like a creature that attacks you, then randomly wanders off. It’s not efficient design, but it creates moments where the world feels alive, or at least strange. That’s something I want to explore more in the future.
Are there any unresolved bugs or limitations you're still dealing with?
Yes, unfortunately. Most of the major launch bugs have been fixed, but there are still a few lingering ones that I haven’t been able to solve on my own.
For example, some Windows 11 users can’t save their game. The OS randomly blocks the game from writing to disk, maybe because it sees it as unverified software. Then there’s lag-based disconnections—players from very distant regions sometimes get auto-kicked by the server due to Unreal’s internal networking limits. And finally, some very old or unusual GPUs crash the game when picking up objects. I suspect it’s a shader issue, but I haven’t been able to reproduce it reliably.
These are edge cases, but they still bother me—especially because part of the goal with Flyknight was to make a game that could run on lower-end machines. I hope I can eventually get help from someone more experienced with low-level debugging, but for now, those bugs are still on the list.
What are your plans for Flyknight going forward?
Even though Flyknight started as a small detour, it’s definitely not over. I’ve already announced a few updates I want to make, starting with a New Game+ mode that would change how you experience a second playthrough. I’m also working on a randomizer mode, inspired by Dark Souls randomizers I’ve enjoyed over the years—where everything from loot to enemy placement is reshuffled.
Beyond that, I want to localize the game into more languages—it’s already available in Spanish, and setting up the localization system took some time, but now it’s ready for expansion. And yes, I’d love to bring it to consoles, though I’m hoping to partner with someone else to handle the porting. Doing it myself, especially with the whole game in Blueprints, sounds like a nightmare.
Will there be a sequel?
Yes, we want to make a sequel. The world of Flyknight—this tiny, humanized insect society—turned out to be incredibly fun to build. It sparks a lot of imagination, both for us and for the players. People keep asking about a sequel or DLC, and we’ve already started planning ideas.
That said, we’re not jumping straight into it. I got a bit burnt out during the final stretch of development—especially in January and February when I was crunching to fix bugs after release—so for now, I’m returning to work on Sorceress, which has been on hold.
