Unity Heart of the Tower [Dev Thread]

[Sentinel Sparrow]

here's a concept of the bunny maid, it took all day.
I did try an ai modified version. There's definitely consistency limitations for me.

Also! 10 extra ai modified bunny maid images

You must be registered to see the links

.

Happy Easter ✨ ^ w ^

 

[DuckHeal]

I might be old-fashioned, but I like the hand-drawn version more...

Edit: looking more closely, the AI seems to lose some little details here and there

 

[Sentinel Sparrow]

Heyy, formerly NMMMN here. I've officially rebranded under the name Sentinel Sparrow, my artist/dev identity moving forward. Why the change? Because I want to create with more freedom — to be fully myself, without the pressure of being seen as an official studio.


Your support fuels the journey — and the journey is still focused on bringing Heart of the Tower to life with everything I've got.


The final game will be published under Stellar North Games, but here? You’re supporting me — the artist behind the fire.



Don’t worry —I’m still the indecisive lovable mess you patiently guide with good reasoning.

—Sparrow

 

[Sentinel Sparrow]

the AI seems to lose some little details here and there

Absotutely! Still working on my asset workflow. I've been gauging how long it takes and its looking to be about 1 week for a character animated in one pose. Also looking at game systems that i can make upfront to then leverage to allow for more focuse on monthly content dev cycle (boobs animated / levels / etc).

 

[Sentinel Sparrow]

The intro scene is drafted!
Guess im writing a book rn lol. Not sure if i should post it on patreon, maybe post monthly chapters or something.


I have the end goal / scene planed out.
I need to figure out the 8 goddesses arcs =.=" im thinking of a narrative loop that i can re-use for them, but also like the progress of just doing it writing and figuring it out but then it might not flow, character missalignment or plot holes and so on /=.="

im still gunning for the 0.0.1 proto demo lol, once i have the story arc and characters figured out, i can then do a full character concept to pose to animation then can do the facial animation gesture system. And the intro sequence will take a bit to animate, though im trying to make the writing speak more than the visuals can give (to save time animating and bc my art style is kinda lacking so far).

I really wanted to hit a demo for Steam Next fest June 1st have the page setup. Apparently its good to have it done 2 weeks before, so 1 week for approval and 1 week for early press access. Spoiler i'll be lucky if i have the main story arc figured out.

luckily i can start saving on my mint tea now that my gardens growing ^ ^;;

 

[Sentinel Sparrow]

I'm actually 3/4th done the intro ^ ^; few things left to change.
My focus shifted to researching AI workflows, AI copyright and testing

You must be registered to see the links

image to video AI that's p cool.

Spoiler: FramePack info

You don't have permission to view the spoiler content. Log in or register now.

new version of my bunny goddess. still not good enough for me, not my style.
masterdebating workflow , might do 3D or stick with my 2D workflow



FramePack has nice dynamic movement. The problems are its hard to get the desired movement of everything, fast movement causes pixellation. I might just make 2D puppet rigs in unity, maybe using framepack clips as ref pose...

View attachment 19 - bunny goddess turn video 5-18-2025.webm
View attachment 17 - bunny goddess video 5-18-2025.webm

 

[godlovesyou]

Sentinel Sparrow

What's your plan for gameplay with all of this?

 

[Sentinel Sparrow]

Sentinel Sparrow

What's your plan for gameplay with all of this?

Core Game Loop

• First-Person Perspective Combat:
  • Melee focused: light and heavy attacks, parry, block
  • Multiple weapon types (e.g. sword, axe, claw, etc.)
  • Visceral, close-range encounters
• Loot & Progression:
  • Find weapons/armor (Diablo 2-style random loot generation)
  • Each run has the potential for powerful finds
  • Extraction-style survival: Find an exit
• Death System:
  • Permadeath-lite: On death, one chance to reclaim your body (and loot/xp) from where you died
  • Failure to retrieve = xp and any gear carried loss (like Escape from Tarkov or Dark Souls soul recovery)
• Return to Sanctuary:
  • Use XP to level
  • Store or upgrade gear
  • Access narrative, romance, and sex scenes with goddesses
  • Spend favor on which goddess you want to progress and earn buffs from

Goddess System

• Visual Novel-style Story Arcs:
  • Each goddess has a linear romance arc with some branching/choice scenes
  • Heavy emotional themes with intimate, soul-connected tone
  • Sex scenes: 3 defined poses, 3 speed variations, climax animations
• Found via Tower Progression:
  • Some are discovered linearly, based on story progression
  • Some are hidden/optional, found through exploration
• Sanctuary Hub:
  • A place for goddess interaction, rest, upgrades, dialogue, romance
  • Central loop node: Combat → Sanctuary → VN progress → Back into the Tower

Let me know if there's more aspects you'd like to learn ^-^

 

[Sentinel Sparrow]

Sentinel Sparrow

What's your plan for gameplay with all of this?

im revising the games end goal. High level goal: 2 year project, concise game like D2 LOD, linear with replay builds.
setting up

You must be registered to see the links

to manage GDD, characters and track progress.
Looking into speech to text to speed up my workflow,

You must be registered to see the links

is promising but i need it to work with GPT but im not sure how.

 

[Sentinel Sparrow]

Thank you for checking in godlovesyou ^ ^; That's a really good question btw.
I kinda got lost in the sauce the past 2 weeks. Once I clarify my end goal and setup GPT projects ill be able to maintain focus. Then my determination will resolve the rest.

 

[godlovesyou]

Core Game Loop

• First-Person Perspective Combat:
  • Melee focused: light and heavy attacks, parry, block
  • Multiple weapon types (e.g. sword, axe, claw, etc.)
  • Visceral, close-range encounters
• Loot & Progression:
  • Find weapons/armor (Diablo 2-style random loot generation)
  • Each run has the potential for powerful finds
  • Extraction-style survival: Find an exit
• Death System:
  • Permadeath-lite: On death, one chance to reclaim your body (and loot/xp) from where you died
  • Failure to retrieve = xp and any gear carried loss (like Escape from Tarkov or Dark Souls soul recovery)
• Return to Sanctuary:
  • Use XP to level
  • Store or upgrade gear
  • Access narrative, romance, and sex scenes with goddesses
  • Spend favor on which goddess you want to progress and earn buffs from

Goddess System

• Visual Novel-style Story Arcs:
  • Each goddess has a linear romance arc with some branching/choice scenes
  • Heavy emotional themes with intimate, soul-connected tone
  • Sex scenes: 3 defined poses, 3 speed variations, climax animations
• Found via Tower Progression:
  • Some are discovered linearly, based on story progression
  • Some are hidden/optional, found through exploration
• Sanctuary Hub:
  • A place for goddess interaction, rest, upgrades, dialogue, romance
  • Central loop node: Combat → Sanctuary → VN progress → Back into the Tower

Let me know if there's more aspects you'd like to learn ^-^

Cool, I love diablo. Are we going to play as the goddesses or guys?

So, is it gonna be like you a girl seducing dudes?

Or like, you are a guy doing missions for the goddesses?

What about outfits?

 

[Sentinel Sparrow]

excellent questions

The D2 part is more about the itemization / gear you find, and linear story act structure, it might feel more like the new doom game which is funny bc its focused on melee combat. Which i didnt even know till recently, everythings been done before.

oh no you'll be playing as a First person perspective of dude man bro. Wimpy chap to giga chad, aiming for a MC personality like Dante DMC / Inu Yasha / Ryudo. His relationships with the goddesses are hetero.

good question i havent really thought about the relationship dynamics other than personality types and sex abilities for the goddesses. So more like fetch quests / relationship arcs. As i build the story i'll know more

So your mission is to get to the choppa, uh i mean to the top of the tower.
I think "narratively" you'll be unlocking goddesses from altars you find along the way, though dont hold me to that one.

outfits for the ladies is a nice to have, not likely atm. maybe dlc

 

[Sentinel Sparrow]

My brain is scrambled since the release of FramePack and learning about img to video, img to 3D.
I cant decide how im gonna make the art. It's p bad lol.

There's so many options. I need to make a character concept to 1 idle pose and 3 sex poses in at least 1 week so i can develop the game. With 8 goddesses at 1 per month thats 8 months. But i think i can only manage 1 animated pose in a week. =.=" i'm so disheartened.

1st step is certain: sketching to Ai for authorship of character IP. I do want to release on steam, own my characters and make them visually consistent.

The next steps goes into 3 ways, the goal is to have a final pose ready to animate.

Methods:

1. sketch it
   Drawing every pose is direct and simple but inefficient over time.
2. make a 3D model / rig.
   Large upfront setup time cost. Faster animation with body and face mocap, physics simulation.
3. use Dzine ai to pose characters
   Faster but results will need alot of editing. Token cost.

3D/2D hybrid workflow idea:

1. 3D generic female character model (no features)
2. Pose the 3D model
3. Sketch over this with the characteristic
4. Ai style and color using previous image as input
5. Cut the image into parts for a 2D puppet rig
6. save PSB file in unity
7. create 2D rig with unity
8. then animate

optionally:
*use FramePack to animate parts: hair, face expression, clothes, particle fx. then save them as sprite sheets.
*use FramPack to rotate character, or get next keyframe, sketching/Ai style and editing the result as needed.


And this is just for the characters...
was thinking enemies would be 2D sprites / rotate towards camera, but now im thinking img to Mesh, same for props and weapons. Some of these ai models require bigger GPU's and the market is going up ;-; 3090 refurbished is going for 2k CAD and the 1.2k workstation to utilizie this beast of a gpu. i really like to buy one but the cost puts me a serious risk literally maxing my credit i would have to earn 2k within a month and that aint happening based on my squirrel brain that cant decide and get shit done and the stats of patreon earnings within the year... =.=" the new workstation will be a huge upgrade once i buy it, it feels like a necessity more than an upgrade.

High level list of to do's

• organize notes / gpt projects setup?
• define end goal / scope
• define asset workflow and try it
• decide on art style
• finish writing intro
• start coding prototype
• try not to feel overwhelmed

 

[DuckHeal]

Perhaps, you could start with a prototype. In game dev, the code usually comes first. Code is like bones in body, if you have a solid skeleton you can pull whatever art onto it. If you don't have a skeleton, nothing will hold, flesh will crumble into a pile.

There's also an issue of code evolving and changing as you go, especially if you're a beginner and/or a single dev who doesn't have a manager with a big stick standing over him and is free to change the whole game design at the whim of his right foot. So you can potentially end up with a pile of redundant (like, you didn't know you could reuse a certain art piece later in the game, because you had no idea you'd decide to add this code to the game later, so now you have two piece of art doing essentially the same thing, double work), incompatible and unused art, while missing critical art pieces that you really need

 

[Sentinel Sparrow]

Perhaps, you could start with a prototype. In game dev, the code usually comes first. Code is like bones in body, if you have a solid skeleton you can pull whatever art onto it. If you don't have a skeleton, nothing will hold, flesh will crumble into a pile.

There's also an issue of code evolving and changing as you go, especially if you're a beginner and/or a single dev who doesn't have a manager with a big stick standing over him and is free to change the whole game design at the whim of his right foot. So you can potentially end up with a pile of redundant (like, you didn't know you could reuse a certain art piece later in the game, because you had no idea you'd decide to add this code to the game later, so now you have two piece of art doing essentially the same thing, double work), incompatible and unused art, while missing critical art pieces that you really need

yeah the project will evolve, i understand that prototyping the gameplay is a good way to understand the needs. I'm just of the mindset that once i know exactly what im doing i can develop the game with focus and consistent updates. because i know im slow and have a hard time deciding, I want to mitigate this by figuring out as much as i can upfront so when i have a prototype and the games rolling im not stuck weeks months trying to decide on things while my patreons expect consistent updates.

If I can front-load the vision, I’ll reduce future drift and delays for my supporters. Thus im looking into obsidian and notion for the high level project management so i dont loose focus and can track ideas / progress. I've tried projects in GPT plus and it feels like any other thread with more context, more time spent guiding GPT to make documents correctly rather than just doing it myself in google docs. google docs feels like i need a high level view of it thats why notion sounds appealing but i dont want to fall into the trap of building the perfect tracker app, same feeling for obsidian but i think with windsurf there's potential for story boarding, but im skeptical about using it for a GDD.

maybe you're right, i just need to make the thing instead of thinking about it. idk there's a balance.

 

[Sentinel Sparrow]

i've been thinking... maybe ill make something different. What I can do well and relatively with ease is 2D game like What a Legend kind of scenes, static camera. Heart Of The Tower would require 3D modelling for the weapons, environment possibly characters. I tried Meshy.ai 2.5 and it produces rough sculpts but the texture and UV's are terrible. It cant be re-textured bc of the patchwork UV's would cause seems, also the texture is regenerated ai shlop. better off learning 3D.

here is a comparison of the swords :
1) texture vertex lit

2) POM shader (looks better bc of metallic, smoothness settings on the shader, also the Parallax occlusion mapping is set to be raised slightly giving it the tiniest bit of depth).

3) 3D meshy ai sword:




So i'm torn, running in circles trying to decide what i should do.
either i learn 3D, or pivot and make a 2D game. For exploring worlds I prefer 3D going through the world not just side view platformer. 2D top down and isometric games have this effect but require a player avatar. 3rd person works well for platformers and souls like, melee combat in first person tends to feel flat. 2D characters would have easier time for outfits than 3D, but 3D can be re-used in any angle but cost more performance.

3D porn game would be more desirable. Ai 2D games will become the norm and it will be hard to stand out like how Unreal Dev's are struggling to make their games look unique bc the engine has a distinct style out of the box.

i've been trying to re-imagine the porn game i would want to play vs what are generally fun game mechanics vs what are good porn game mechanics and im interested in hearing what others think makes a good porn game.

things i'm considering and struggling to decide for the new game or rework of the game
- 1 hand controls (entire game vs only sex scenes)
- sex focused gameplay vs separated
- 2D only gameplay or 3D
- deep relationship, slow burn relationship?
- harem flings would be nice, though sticking to one character especially 3D would be easier, if 2D then a character puppet rig.
... i have a bunch of ideas scrambling my brain.

I worry that this will be the only game i ever make or develop for a long time this why I'm hesitant on making a 2D game that I wouldnt necessarily want to play even though it would be "simpler" to make. everything will take me a long time =.="

to make a game that you go through the world while being 2D would be top down, but it falls flat from above. i tried at an angle but when moving on the z axis passing 2D sprites are paper thin. I've tried a 2D sword in first person but it felt paper thin as well.


here is examples of the sword and building issue that arises from this method.
(i should of animated the sword swinging to better show how thin it looks)
View attachment 21 - angled camera 2D building and tree test 3-6-2025.webm
View attachment 20 - sword test 3-6-2025.webm