🟢 Xbox2017Studio MDHRVERY HARDCuphead rebuild deck
VERY HARD
The data-driven boss phase machine is an afternoon; the game is thousands of hand-inked cel frames and a live big-band recording, so your bottleneck is Aseprite and it is not going to move.
You will get a good prototype. The product is what comes after.
Worth saying plainly, because it is the opposite of everything else on this page: the difficulty here is art and animation, not engineering. An agent hands you a working, data-driven boss machine in an afternoon and then you are standing in front of the real problem, which is thousands of hand-inked frames and a live jazz recording. Spend your review time on the hitbox tables and the parry window, then go and learn to draw, or hire someone who already can.
- Difficulty
- VERY HARD · 4 of 5
- First playable
- A weekend
- The original
- 60 team-years
- Released
- 2017 · Xbox
- Genre
- Run And Gun
Build a run-and-gun boss fight. ONE boss, three phases. Canvas 2D, single page, fixed 60Hz simulation. The engineering here is the smallest on this site; the art is the game and you are not getting the art. Requirements: - Author every hitbox as an explicit rectangle in a data table keyed by animation frame. Never derive a hitbox from sprite bounds or a transparent-pixel scan. The player's hurtbox is a fixed rect noticeably smaller than the drawn character and does not change across their whole animation set, and every projectile carries its own hitbox independent of its art. This single decision is the difference between a fair bullet-pattern fight and an infuriating one, and animation-driven games get it wrong constantly. - The boss is a deterministic phase machine defined in JSON: each phase holds a weighted list of attack patterns with fixed telegraph, active and recovery durations in ticks, and never repeats a pattern twice consecutively. - Shooting is eight-directional and independent of facing, with a hold-to-lock-aim modifier and a dash with a short invulnerable window. - Three hit points, no i-frames beyond a brief blink, no healing. - Parry: objects flagged parryable are cleared by pressing jump a second time in the air, granting meter and an upward hop. - Death screen shows how far into the fight you reached and the name of the pattern that killed you. Definition of done: phase two kills me at 80%, I retry in under a second, and I know exactly which pattern did it.
- The art, which is the entire game — every frame drawn, inked and painted by hand in a 1930s cel workflow, over seven years, by people who had to learn to do it
- Watercolour backgrounds painted as single pieces rather than assembled from tiles
- A live-recorded big-band jazz score that carries the pacing of every fight
- Boss patterns iterated for years — the code for a phase machine is an afternoon, the patterns worth putting in it are not
The parry — designated hazards become a resource if you touch them on purpose
It inverts the meaning of a threat with one flag on an object, and it works anywhere there are incoming hazards: a shmup, a rhythm game, a tower defence, a fighting game. The design cost is a reserved colour and a timing window.
ScopeA weekend— build just this, not the game
Single self-contained HTML file, Canvas 2D. A platformer floor, one stationary emitter firing projectiles in patterns, a player who can run, jump, double jump and parry. - One projectile in three is flagged parryable and drawn in a single colour reserved for nothing else in the entire toy. That colour is a contract with the player — never reuse it for decoration. - Parry input is the jump button pressed a second time while airborne. Overlapping a parryable object during the window destroys it, grants a small upward hop that refreshes the double jump, and adds one unit to a five-unit meter. - A missed parry costs you: it consumes your air action, so mistiming leaves you falling with nothing left. - At five units, a separate button spends the whole meter on a screen-clearing attack. - Hitboxes are explicit rects from a table, never sprite bounds. Make the parry window forgiving — six frames — while the damage hitbox stays tight. Definition of done: I cross a stream of projectiles by choosing to touch three of them.
Not a clone — Cuphead re-themed and cut down to something you can finish. The angle is the important reel: it is what turns a project into a weekend. 5,880 ways to land.
Deniable Hands Off
Cuphead, but an auto-battler — you set up, it plays itself, set in Cold War espionage. Rendered in Game Boy four-shade green.
Build a browser game called "Deniable Hands Off". The pitch in one line: a run and gun in the spirit of Cuphead, but an auto-battler — you set up, it plays itself, set in Cold War espionage. WHAT THIS IS Take the SHAPE of Cuphead — its core loop, and whatever makes that loop good — and rebuild it as your own game in a different world at a smaller scope. Use your own names, art and audio, which you want anyway: a reskin of somebody else's game is less interesting than a new one that works the same way. THE ANGLE — this is the important constraint, honour it above everything else An auto-battler — you set up, it plays itself. The player only makes decisions between rounds. Combat resolves deterministically from board state and seed, so a replay is exact. This is what makes the project finishable. If a decision would push the scope back towards the original's, take the smaller option every time. THE THEME Cold War espionage. Nobody can be trusted including your own side, and the failure state is exposure rather than death. The theme is not a coat of paint. Let it change what the mechanics mean — a reload, a health pack and a locked door should all be things that make sense in this world and nowhere else. LOOK Game Boy four-shade green. A strict four-colour palette, 160×144 internal resolution scaled up with nearest-neighbour, and no anti-aliasing anywhere. REQUIREMENTS - Runs in a browser with no build step. Canvas 2D unless the angle genuinely requires 3D, in which case three.js. - Fixed-timestep update loop with interpolated rendering, so it behaves the same at 60Hz and 144Hz. - Every tuning constant in one CONFIG object at the top, exposed as live sliders in a debug panel. You will find the feel by dragging those, not by prompting. - All content — levels, entities, balance numbers — in JSON or plain data files, never inline in the logic. - Audio generated with the Web Audio API rather than asset files. - Respect prefers-reduced-motion: keep the fades, drop the shake and the parallax. BUILD ORDER 1. The core loop with placeholder rectangles, at the reduced scope the angle demands. No theme, no art, no audio. Make it fun as rectangles first. 2. The angle's consequences. Cutting to bots, or to turn-based, or to one level changes the design rather than just shrinking it — find out how and rewrite step 1. 3. The theme, as content and data. 4. The look, applied last as a rendering layer over a game that already works. Start by writing a one-page plan: the core loop, what the angle removes and what that frees you to do properly, and what you are deliberately leaving out. Wait for me to approve it before writing code.
- Aseprite — the animation tool, which is the actual bottleneck here
- OpenGameArt — placeholder sprites while you face facts
Half the time one of these already does what you wanted, and the other half it saves you a week of solving a problem someone documented in 1997.
One count per game. No account — it is tied to a salted hash of your IP, which is not stored in a form anyone can reverse. Every build adds 60 team-years to the counter on the homepage.
Then try these
Questions
How hard is Cuphead to rebuild with an AI agent?
Yes — every game is buildable at some scope. The question is how hard, and Cuphead is VERY HARD (4 of 5). You will get a good prototype. The product is what comes after. The data-driven boss phase machine is an afternoon; the game is thousands of hand-inked cel frames and a live big-band recording, so your bottleneck is Aseprite and it is not going to move. Reckon on a weekend to something you can actually play — not to something that matches the 60 team-years Studio MDHR spent.
What makes Cuphead hard to rebuild?
The data-driven boss phase machine is an afternoon; the game is thousands of hand-inked cel frames and a live big-band recording, so your bottleneck is Aseprite and it is not going to move. That is why it sits at VERY HARD rather than a tier either side of it. If you want the idea without the project, the transferable part is the parry — designated hazards become a resource if you touch them on purpose — a weekend of work.
How long does it take to build a Cuphead clone?
Wrong question, and it is worth saying why. Nobody rebuilds Cuphead — the original took Studio MDHR on the order of 60 team-years, released in 2017 on Xbox. What you can do is reach something playable and genuinely yours in about a weekend, and then keep going for as long as it stays interesting. Any figure that claims otherwise is measuring a prototype and calling it a game.
What do I lose by building my own Cuphead?
The art, which is the entire game — every frame drawn, inked and painted by hand in a 1930s cel workflow, over seven years, by people who had to learn to do it. Watercolour backgrounds painted as single pieces rather than assembled from tiles. A live-recorded big-band jazz score that carries the pacing of every fight. Those are the parts a prompt will not hand you.
Which AI coding agent should I use to build it?
Any of Claude Code, Codex or Cursor will handle this prompt. Paste it as your first message, let the agent scaffold the project, then iterate on feel — the second, third and tenth prompts are where a game actually gets good. Work in an empty folder or a fresh git branch so you can throw it away cheaply.
One email when new games go on the list. No other reason to email you, ever.