Summer Nights - 3D Godot Game
Summer Nights -
3D Godot Game
Summer Nights - 3D Godot
Game
English
Timeline
2 Months +
Role
Design Engineer
Team
2x Designers
Domain
Gamification
Timeline
2 Months +
Role
Design Engineer
Team
2x Designers
Domain
Gamification
Tools & Stack
Figma
GitHub
Antigravity
All project data & decisions documented here are my own. AI assistance was only used for structure and clarity.
All project data & decisions documented here are my own. AI assistance was only used for structure and clarity.
01Business Context & Competitive Landscape
The project originated from a July design prompt in a Kakao Design Club founded by London-based UI/UX designer Yodi (요디님). What started as a weekend concept for a summer theme quickly highlighted a visible gap in indie game development. Many vibe-coded or one-shot game demos fail once a player actually plays them because they lack real UX, rely on stock AI art, and have poorly structured interfaces.
The goal with Summer Nights was to build a finished product rather than a novelty. It is a game with a rigorous design system, zero generative AI art, and deep interactive systems that proves product design methodologies can directly translate to game engines.01Business Context & Competitive Landscape
The project originated from a July design prompt in a Kakao Design Club founded by London-based UI/UX designer Yodi (요디님). What started as a weekend concept for a summer theme quickly highlighted a visible gap in indie game development. Many vibe-coded or one-shot game demos fail once a player actually plays them because they lack real UX, rely on stock AI art, and have poorly structured interfaces.
The goal with Summer Nights was to build a finished product rather than a novelty. It is a game with a rigorous design system, zero generative AI art, and deep interactive systems that proves product design methodologies can directly translate to game engines.02Problem
Drawing interfaces in Figma is easy, but making them feel tactile and scalable in a real-time 3D engine is incredibly difficult. 02Problem
Drawing interfaces in Figma is easy, but making them feel tactile and scalable in a real-time 3D engine is incredibly difficult. Business Problem
The indie game market is saturated with "vibe-coded" demos that fail to retain players because they lack foundational UX and rely on low-quality generative AI, resulting in products that look good in screenshots but fail in actual gameplay.User Problem
Players expect immediate feedback and intuitive interfaces, but most indie games treat UI as an afterthought, leading to high cognitive load and poor user retention.Business Problem
The indie game market is saturated with "vibe-coded" demos that fail to retain players because they lack foundational UX and rely on low-quality generative AI, resulting in products that look good in screenshots but fail in actual gameplay.User Problem
Players expect immediate feedback and intuitive interfaces, but most indie games treat UI as an afterthought, leading to high cognitive load and poor user retention.03My Role
Serving as director, game designer, product manager, and technical UI developer, I owned end-to-end design and implementation. I collaborated with an AI coding assistant (Google Antigravity) for layout math and shaders, and UI artist Ivy for custom branding and vectors.01Full UX/UI Design & Engineering: Conceptualized in Figma, programmed in GDScript in Godot 4.
02Design System Architecture: Built a scalable token system for typography, colors, and components.
03Bilingual Localization: Architected English and Korean (한국어) parity from Day 1.
04Documentation as Truth: Maintained 4 core synchronized doc files to guide all structural decisions.
03My Role
Serving as director, game designer, product manager, and technical UI developer, I owned end-to-end design and implementation. I collaborated with an AI coding assistant (Google Antigravity) for layout math and shaders, and UI artist Ivy for custom branding and vectors.01Full UX/UI Design & Engineering: Conceptualized in Figma, programmed in GDScript in Godot 4.
02Design System Architecture: Built a scalable token system for typography, colors, and components.
03Bilingual Localization: Architected English and Korean (한국어) parity from Day 1.
04Documentation as Truth: Maintained 4 core synchronized doc files to guide all structural decisions.
04 Direction
I replaced the standard approach of hardcoding game UI with a systematic product approach. I enforced a Zero Generative AI for Art policy, relying entirely on Godot StyleBoxFlat vectors, math geometry, custom GLSL shaders, and Ivy's UI art. We also committed to bilingual updates. HUD labels, menus, and changelogs update together in both languages for every release.04 Direction
I replaced the standard approach of hardcoding game UI with a systematic product approach. I enforced a Zero Generative AI for Art policy, relying entirely on Godot StyleBoxFlat vectors, math geometry, custom GLSL shaders, and Ivy's UI art. We also committed to bilingual updates. HUD labels, menus, and changelogs update together in both languages for every release.05Impact
Summer Nights successfully proves that rigorous product design methodologies can be applied to interactive engines.05Impact
Summer Nights successfully proves that rigorous product design methodologies can be applied to interactive engines.Public ReleaseReleasedDaily improvements, weekly full releases
2,770+Views on DC InsideEarned "개념글" (Best Post) with 50+ organic community comments.
100%Code-Driven UIZero static overlays; everything from layout anchors to health bars is scalable.
Public ReleaseReleasedDaily improvements, weekly full releases
2,770+Views on DC InsideEarned "개념글" (Best Post) with 50+ organic community comments.
100%Code-Driven UIZero static overlays; everything from layout anchors to health bars is scalable.
06Approach and Rationale
To prevent visual drift across rapid updates, I implemented strict design rules. I established a design system that centralized visual tokens, including specific colors like Cyber Gold (#F5D76E) and Dark Background (#0D051A), alongside consistent 6px and 8px corner radii. I enforced an even-number font size rule (12px, 14px, 16px, 18px) to keep pixel rendering crisp on high-DPI screens. Finally, I maintained a tight user feedback loop where direct telemetry and community requests drove updates, allowing me to engineer features like vaporwave tracks, retro startup screens, and controller rumble based on actual player demand.06Approach and Rationale
To prevent visual drift across rapid updates, I implemented strict design rules. I established a design system that centralized visual tokens, including specific colors like Cyber Gold (#F5D76E) and Dark Background (#0D051A), alongside consistent 6px and 8px corner radii. I enforced an even-number font size rule (12px, 14px, 16px, 18px) to keep pixel rendering crisp on high-DPI screens. Finally, I maintained a tight user feedback loop where direct telemetry and community requests drove updates, allowing me to engineer features like vaporwave tracks, retro startup screens, and controller rumble based on actual player demand.07Final Solution
The final development loop bypassed traditional slow prototyping. The workflow became: Idea ➔ AI-Assisted Architecture ➔ Headless Engine Check ➔ Live Playtest ➔ Documentation Sync.
This allowed the game to evolve rapidly through 5 phases: from a core loop prototype to combat depth, release engineering, rogue-lite perks, and custom post-processing retro shaders.07Final Solution
The final development loop bypassed traditional slow prototyping. The workflow became: Idea ➔ AI-Assisted Architecture ➔ Headless Engine Check ➔ Live Playtest ➔ Documentation Sync.
This allowed the game to evolve rapidly through 5 phases: from a core loop prototype to combat depth, release engineering, rogue-lite perks, and custom post-processing retro shaders.
08 — Outcomes: What Worked
A list of pointers that worked.01Constraints Spark Creativity: Sticking to low-poly models, flat vector panels, and a fixed palette kept the game cohesive.
02Bilingual Design from Day 2: Forcing the UI panels to handle text expansion and Korean font swaps early saved massive retrofitting costs later.
03Rapid Iteration Outpaces Spec Work: Testing a mechanic or UI element directly in Godot within minutes kept development fast and focused over 60 days.
Limitations & Second-Order Effects
A list of things that I realised.01Engine Learning Curve: Translating Figma's Auto-Layout mental model into Godot's Control Node anchors required significant upfront technical learning.
02Scope Discipline: Certain complex features had to be simplified to maintain development velocity as a solo developer.
08 — Outcomes: What Worked
A list of pointers that worked.01Constraints Spark Creativity: Sticking to low-poly models, flat vector panels, and a fixed palette kept the game cohesive.
02Bilingual Design from Day 2: Forcing the UI panels to handle text expansion and Korean font swaps early saved massive retrofitting costs later.
03Rapid Iteration Outpaces Spec Work: Testing a mechanic or UI element directly in Godot within minutes kept development fast and focused over 60 days.
Limitations & Second-Order Effects
A list of things that I realised.01Engine Learning Curve: Translating Figma's Auto-Layout mental model into Godot's Control Node anchors required significant upfront technical learning.
02Scope Discipline: Certain complex features had to be simplified to maintain development velocity as a solo developer.
09 — Reflections: What I'd Do Differently
Code-First Tokens: Building the design system midway through development created retrofitting costs. The lesson is clear: establish design tokens (fonts, colors, spacing) in the code architecture on Day One before drawing a single screen. We lost time refactoring hardcoded values that could have been avoided with a stricter initial system.What I Learned
Ship and Learn: Being the sole designer/developer means learning to ship directionally right, then refine. The biggest lesson from Summer Nights is the monumental difference between drawing a UI in Figma and making it "feel" good in the hands of a user. The willingness to dive into the code, manipulate state, and wrestle with engine constraints is what separates a static mockup from a living product.09 — Reflections: What I'd Do Differently
Code-First Tokens: Building the design system midway through development created retrofitting costs. The lesson is clear: establish design tokens (fonts, colors, spacing) in the code architecture on Day One before drawing a single screen. We lost time refactoring hardcoded values that could have been avoided with a stricter initial system.What I Learned
Ship and Learn: Being the sole designer/developer means learning to ship directionally right, then refine. The biggest lesson from Summer Nights is the monumental difference between drawing a UI in Figma and making it "feel" good in the hands of a user. The willingness to dive into the code, manipulate state, and wrestle with engine constraints is what separates a static mockup from a living product.