Foreword

i've been wanting to write something like this for a while, not because i think everything we've done is perfect, but because i think the real story of how DELTARUNE: Tower Defense got made is more interesting than anything i could dress up for a press release. there were mistakes, there were moments where we almost stopped, there were bugs that made everyone in the test session lose it completely, and there were decisions that seemed completely obvious at the time and are now genuinely hard to explain. i want to cover all of that, because that's the part that actually matters. anyone can write a list of features. what i want to write is an honest account of what we were thinking, why we made the choices we made, and what the whole thing taught us along the way.

Introducing Myself

hi everyone! you might know me as SnowzBunz, or maybe just Bunz! i'm the Vice President and Creative Director of Lucky 7 Entertainment, and the creator of DELTARUNE: Tower Defense. i've been on Roblox since 2016, and the reason i stuck around long enough to start making games is the same reason a lot of developers end up where they are: i kept playing games that frustrated me, not because they were hard, but because they felt careless. like the people who made them had a good idea somewhere at the beginning and stopped caring before it was finished. that bothered me a lot, and it still does. my role at Lucky 7 means overseeing the creative direction of everything we make, making design decisions, and keeping the team moving in the same direction. since 2024 i've been spending around four to six hours every single day on DRTD, and the rest of the team is putting in roughly the same. one misconception i hear constantly is that game development is simple. people underestimate the conceptual work: deciding what the game is, how every mechanic interacts with every other mechanic, and questioning all of those decisions again every time something doesn't work. the building is almost the easy part.

Lucky 7 Entertainment

Lucky 7 Entertainment was founded by me and Luci with one goal: we wanted to make Roblox games that people wouldn't believe existed on Roblox. High Stakes Vegas pushes into territory most developers wouldn't touch on the platform, mature gameplay concepts similar to what you'd find in GTA 5, built inside Roblox. DELTARUNE: Tower Defense is a completely different kind of attempt, but the underlying belief is the same. the platform has more capacity than most people give it credit for. before DRTD became our flagship project, we were developing Zombie Cells, High Stakes Vegas, and Battleship RIMPAC alongside each other. each of those taught us something that eventually shaped how we approached DRTD. there was also a project called Chernobyl, a survival game built around the nuclear disaster concept, which was cancelled before players ever saw anything. we plan to come back to it someday. the long-term goal for Lucky 7 is to finish DRTD, then shift full focus to High Stakes Vegas and Battleship RIMPAC.

Before DELTARUNE: Tower Defense

the version of me that started working on DRTD was someone who had been playing Roblox tower defense games for a long time and was increasingly frustrated with what they were doing. that frustration had a specific focal point: Five Nights TD. i was a top-three leaderboard player in FNTD, grinding it constantly. the game had something real in it, the tower defense format worked with the FNAF universe in ways that felt natural. but the longer i played, the more obvious it became that the care players were giving the game wasn't being matched on the development side. that planted a specific question in my head: what would this kind of game look like if the people building it genuinely treated it like it mattered? of all the genres we could have chosen, tower defense allows the most creative freedom. the core loop is stable enough to hold almost any mechanic, but specific enough that you always have a frame to work inside. units can have completely different abilities. maps can have different layouts and rules. events can introduce mechanics that don't exist anywhere else in the game. the genre bends without breaking.

Choosing DELTARUNE

i've been a DELTARUNE and UNDERTALE fan since around 2016. i used to watch gameplay videos and play roleplay games based on the universe constantly, and alongside FNAF it was probably the thing i cared most about at that point. when DELTARUNE started releasing chapters, i followed it the same way i'd followed UNDERTALE: completely. the practical reason DELTARUNE made sense for a tower defense game is that its characters already have combat roles built into them. Kris has a sword, Susie has an axe, Noelle has ice magic, Ralsei heals. they're RPG party members, so assigning them to tower defense mechanics isn't a stretch, it's almost natural. the Dark World setting also gives you enormous artistic range: Castle Town to Queen's palace to the Snow Grave, all without it ever feeling like a different game. at the time we started, there wasn't a definitive DELTARUNE tower defense game on Roblox. that's a specific kind of opportunity. the game also grew alongside DELTARUNE itself: when Toby Fox released Chapters 3 and 4, we got Lighteners Live, which we turned into a fully working karaoke mode, and new mechanics like Darkness. Chapter 5 gave us the concept for the Deity tier.

The First Concepts

the earliest version of DRTD was a collection of MS Paint sketches. rectangular boxes with character names next to them, arrows indicating attack range, notes about what abilities might look like. nothing that would communicate anything to anyone who wasn't already inside my head imagining the same thing. the name came before almost everything else. we had DELTARUNE: Tower Defense as a title before we had a single working unit or map, and that actually helped because it gave the project direction before it had a shape. knowing it was a DELTARUNE tower defense game meant every early decision had a frame to operate inside. the first unit to get full stats and a proper concept was Darkworld Kris. what's interesting is that Kris has stayed largely the same since early 2025, while most units have gone through significant revision. one early mechanic we explored was the ability to play as units directly during a wave, but when you give players direct control of one unit they stop thinking about the whole map. the strategic layer that makes tower defense interesting collapses. it was cut, and that was the right call.

Building the Team

i'm not a modeler in the way the game needed. i use CAD-based software like SolidWorks, an engineering tool, not the kind of organic low-poly character work a Roblox game requires. so when the concepts started solidifying, i posted an ad on HiddenDevs looking for a low-poly modeler. that's how we found Gladiator, whose real name is Taimoor and who is now our Modeling Director. all 306 units in the game are his work, built from scratch from my concepts. low poly isn't a limitation in his hands, it's a language, and he's gotten very good at saying things in it. Art Direction came from Alexis, our Art Director. the first thumbnail she made was for Susiezilla and the Delta Warriors, and it immediately set a tone for the whole game. Alexis later recommended her partner Narrat0r_ to share the load of 20+ thumbnails and 40+ badges. Noei joined as Audio Director and took on the full 150+ track original soundtrack. Gizmo came on as a concept artist, bringing ideas we wouldn't have reached on our own. Luci co-founded Lucky 7 with me. Bekir joined as Lead Advisor. without all of them, there is no DRTD.

The Rebuild

the first version of DRTD got far enough that we scheduled an alpha test. we ran it. it was not good. the texture work was inconsistent, units placed near each other looked like they came from different games. the UI worked technically but didn't feel right. there were bugs that weren't just cosmetic. deciding to rebuild was not easy. we'd put real work into the first version, and the argument for continuing to patch rather than starting over is always available and never quite honest. we scrapped everything and rebuilt from scratch. every unit was remodeled. the animation and rigging pipeline was rebuilt to a new standard. the UI system was rewritten entirely. and in the process, the game became significantly more ambitious. the Christmas event, the Halloween event, and more than ten new modes all came out of the rebuild period. looking back, the rebuild was the point where DRTD actually became DRTD. the first version was us figuring out the game. the second was us building it.

Designing Units

the roster grew alongside the chapters. when we started with only Chapters 1 and 2, sixty units was roughly the ceiling. every chapter release expanded what was possible. a significant number of units are actually based on community artwork from the DELTARUNE fandom, which is something we're genuinely proud of. every unit in DRTD has unique upgrades. no unit gets generic stat boosts or percentage increases. the balance philosophy reflects the same thinking: an Epic unit used correctly should be capable of outperforming a Deity. we constantly test units against each other in real play conditions to understand how utilization and placement affect actual performance. no unit should be objectively dominant. Omega Flowey was one of the most technically painful units we've ever built. his fake stats caused intermittent parsing errors that were nearly impossible to track down and took weeks to fully solve. he almost got removed. he didn't, and the game is better for it. Kitsune Queen is my personal favorite unit in the game.

Building the Lobby

the original lobby was Castle Town. it made sense at the time: Castle Town is where the main DELTARUNE story takes place and has immediate recognizability for anyone who knows the source material. but Castle Town is a finite space, and when you need to fit every building, every NPC, every mode selector, and all the infrastructure a working lobby requires into a space not designed to hold any of that, you start making compromises. Hometown gave us flexibility Castle Town never could. Alexis rebuilt it from scratch, not as a direct replica, but as a version that could hold everything while still feeling like somewhere a player would want to spend time. every building has a fully modeled exterior and interior. every building has NPCs you can talk to. the decision to give buildings interiors came from one specific goal: we wanted the lobby to feel lived in rather than functional. QC's Diner is my favorite building. Carol's Mansion was the hardest to build. NORMAL_NPC is my favorite NPC in the game.

PvP Philosophy

PvP in Roblox tower defense games has a consistent problem: it's usually not really PvP. it's a comparison of who has the stronger units. whoever spent more time or money wins. the outcome is determined before the match begins. that's not competition. the reference i kept coming back to was Clash Royale, specifically Draft mode. Draft works because it removes the question of who has the better collection entirely. both players draw from the same pool and the result is determined by thinking rather than inventory. that's the philosophy behind DRTD's PvP: the player who wins should be the one who made better decisions, not the one who paid more. Gaster's Lab grew from the same idea. it takes players who believe that having the strongest units means winning, and proves that placement and strategy matter more. it challenges your critical thinking on where to place units and how to utilize them.

Clans

clans exist in DRTD because we wanted players to have a reason to organize with each other that went beyond a shared label. most clan systems on Roblox are essentially cosmetic. what we wanted was something that required actual communication and shared objectives that a single dominant player couldn't carry alone and that automated play couldn't game. the 6+ clan modes prevent macroing and AFK farming from working because each mode requires different kinds of active thinking. some reward strategic unit placement. some reward communication. some reward speed under pressure. a clan with one very powerful player and a dozen inactive members will lose to a clan with twelve engaged players who actually work together. clan units are genuinely exclusive in the sense that you have to play the content to earn them. syncing 50+ players to one clan and keeping accurate track of everything was the biggest technical challenge the system created, but it was worth it.

Music

the most common question i get about the soundtrack is some version of why. why 150 songs, why original music, why invest that much in audio for a Roblox game. the honest answer starts with vinyl records. the long-term vision for DRTD includes physical releases, and a soundtrack pressed to vinyl is something we've wanted since early in the project. we originally used the DELTARUNE soundtrack, but decided we wanted something fully ours. Noei takes Toby Fox's musical sensibility and reinterprets it rather than reproducing it, creating something that feels familiar to DELTARUNE fans without being a copy of what already exists. the count exists because every area has its own music. every map, every event, every corner of the lobby has a track written specifically for it. i hated games that reused one song for hours in the same area. i wanted every part of DRTD to have its own atmosphere and its own emotional register.

Art Direction

the art style of DRTD is something i'd describe as deliberately goofy. DELTARUNE has a specific tone: charming, slightly odd, occasionally unsettling, and never taking itself too seriously even when the story is doing something emotionally heavy. the art that works best in DRTD captures that quality. low poly was the technical choice that made the artistic choice possible. Roblox doesn't perform well with high-poly animated models at the scale DRTD requires. Gladiator has built inside that constraint for years and turned it into the visual identity of the game. the silhouette clarity you get from low-poly work is one of the things that makes the game look the way it does. favorite thumbnail: Delta Warriors. favorite render: Yellow firing into the sunset. favorite badge: Top 250 PvP Winner.

AI NPCs

the NPC conversation system in DRTD is, as far as we know, the first real-time AI integration at this scale on Roblox. most NPC dialogue in Roblox games works on a tree model. players see options, select one, get a scripted response, repeat until the conversation ends. that works, but players who spend enough time with an NPC will exhaust everything it can say. the character stops feeling like a character and starts feeling like a menu. what we built lets players have actual conversations with units in the lobby in real time. there's no dialogue tree. players type what they want to say, and the unit responds based on who that character actually is, their personality, their history in DELTARUNE. if the goal is for the lobby to feel lived in, then the characters who live there have to feel like they actually do.

Hidden Details

the Forgotten Man has eggs scattered throughout various maps and spaces in the game. most players will never find all of them. that's fine. the players who do will find something placed there specifically for them, for the curious ones who are actually looking. that's the kind of player we built DRTD for. every unit in the game uses Nubert as a placeholder. if a unit's model fails to load for any reason, you get Nubert. if any mesh in the game fails to load, it becomes Nubert. it's practical, but it's also become an internal tradition at this point. the unit called ??? almost had to be renamed because Roblox's JSON handling was struggling with the question marks in ways that caused intermittent parsing errors. the solution was giving the unit a separate internal script name while keeping the player-facing name exactly as it was. scrapped mechanics include creating your own units for an event, and gambling with D$. neither of them has been completely forgotten.

Funny Bugs

the most memorable was the weld failure. the welds that keep units rigged are infrastructure you never think about because they always work. until one testing session where every unit simultaneously lost its welds. with a physics engine suddenly having no reason to keep anything in place, units launched across the map, units fell through the floor, units ended up in places that made no structural sense. the session lasted roughly the amount of time it took everyone to stop laughing long enough to restart. Roblox once updated how bones worked without giving developers enough warning. the result was a large number of units having their bone orientations flipped. models that were supposed to have outsides now had insides. the single worst event was a crash during the creation of the main core UI system. over 30,000 lines of work, gone. we rebuilt it. the version that exists now is better than what was lost.

The Community

the thing that surprised me most about DRTD's community wasn't the size. it was the attention. when players make YouTube videos about the hidden details in the game, the Easter eggs, the secrets, the mechanics that only appear under specific conditions, it means they were paying close enough attention to notice things that most players walk past. that's exactly the kind of engagement the game was built for. not the player who clicks through menus. the one who's actually trying to understand what's there. the support during the period when Roblox introduced their new age rating verification requirements meant everything. the new rules required games to either have 500+ highly engaged verified players or pay 100,000 Robux to stay accessible. the community didn't leave during that period. they stayed, kept talking about the game, kept building anticipation for something that still wasn't released. when something you've put years into is threatened by a change you can't control, the people who show up anyway are what make continuing possible.

Looking Toward Release

the current state of development is closer to release than it's ever been. the revamp work is largely finished. the core systems are all in a state i'm genuinely confident in. animation work is being wrapped up. every day the distance between where we are and where players actually have the game is shorter. Mt. Ebott is the thing i'm most excited about. the concept of permanent events, UNDERTALE and Undertale Yellow content available any time rather than during specific seasonal windows, is something we think is genuinely new in this format. Christmas and Halloween are seasonal because that makes sense for them. Mt. Ebott is different. those units and that content should be available to any player who wants them, whenever they want them. the AI NPC system will keep developing. the long-term goal is lobby characters who feel genuinely alive, who have things to say about what's happening in the game. animations are the final major production focus before release. it's close.

Final Thoughts

i've been an UNDERTALE fan since around 2016. i watched gameplay videos and played roleplay games based on the universe before DELTARUNE even existed. alongside FNAF it was probably the thing i cared most about at that point. never at any point during any of that did i think i would turn that obsession into a game that would consume years of my life and the work of an entire team. DRTD is in the simplest possible terms the game i always wanted to play and couldn't find. sixty units became 306. Castle Town became Hometown. scripted dialogue became real-time AI conversations. but the underlying desire never changed: i wanted a tower defense game where every decision felt intentional, where units felt like real characters, where the music made the space feel alive, where players could tell that the people who made it actually cared. development has changed me. the most obvious thing is communication. managing a team across a project like this requires skills i didn't have when we started. every hard period built something i didn't have before. what i hope players remember is the care. not specific features, but the overall sense that someone thought about every part of this and wanted it to feel worth playing. to everyone who stayed, who made videos about the secrets, who told people about the game before it was even out, who never left during the hard times: thank you. this exists because you made it worth existing. thank you for watching our dream.