How to Make a Game Like Slay the Spire, With Your Own Deck and Your Own Climb
Slay the Spire, by Mega Crit, turned a card game into a run. You start with a thin, weak deck, climb a branching map, and grow the deck one choice at a time until it can beat the boss, or until it cannot. Its pull comes from a handful of clear systems, and each of them can be learned and rebuilt in a world of your own.
This guide takes those systems apart, lists what a game in this family needs, then walks through making yours with Tyrbo, the AI game engine, from the first sentence to a signed Android build. The aim is a deckbuilder people remember for its own idea, not one that reminds them of somebody else's.
Card battler and autobattler · Gilded fantasy
Card battler and autobattler · Painted storybook
Card battler and autobattler · Anime bright
Card battler and autobattler · Bright and bold
Card battler and autobattler · Ink sketch
Card battler and autobattler · Spooky cartoon
Card battler and autobattler · Clean neon
Card battler and autobattler · 16-bit pixel
What makes Slay the Spire work
The outer loop fits on a sticky note: fight, take a reward, choose the next room, repeat until the boss. Inside each fight sits a smaller loop. You draw a hand, spend a small pool of energy on the cards you can afford, watch the enemies act, and discard what is left. Nearly every rule in the game exists to make those two loops produce interesting choices.
Energy and the hand: a budget spent every turn
Each turn gives you a fixed amount of energy and a fresh hand drawn from your pile. Cards cost energy, so every hand is a budget puzzle: the expensive attack, or two cheap blocks? Because the hand is random and the budget is tight, no two turns ask quite the same question. Unplayed cards go to the discard, and when the draw pile runs dry the discard is shuffled back in. That cycle is why the makeup of the whole deck matters more than any one card in it.
Enemy intent: a guess becomes a puzzle
Enemies show what they are about to do: attack for a stated amount, defend, weaken you, or something stranger. That single piece of information changes the genre. Instead of guessing, you read the threat and answer it, blocking exactly enough or racing to finish the enemy first. If you carry one idea from this game into yours, carry this one. Hidden information makes a card game feel like luck; shown information makes it feel like skill.
The deck grows, and shrinks, between fights
After a fight you are offered a few cards and may take one or skip them all. Shops sell cards and let you pay to remove one. Rest sites trade healing against upgrading a card. Players slowly learn that a deck is a set of odds: every weak card added makes the strong ones turn up less often. When skipping a reward is a real decision, the economy is working.
Relics: rules that bend for one run
Relics are passive items found along the climb. Each changes a rule until the run ends: an extra card on the first turn, energy when something specific happens, a bonus for a deck that leans one way. Relics give every run a personality and nudge players toward builds they would never have tried on purpose.
The map: choosing your risk
The climb is a branching map of rooms: ordinary fights, tougher elites, shops, rest sites and unknown events, with a boss at the top of each act. Seeing the paths ahead lets you plan. A strong deck hunts the elite for its better reward; a fragile one takes the long way round to a campfire. Strategy happens between fights as much as inside them.
The parts a game like it needs
A deckbuilder breaks down into three kinds of work: systems, content and tuning.
Systems
- The card cycle: draw pile, hand, discard, a pile for cards removed for the rest of a fight, and a reshuffle rule.
- Energy and costs: how much arrives each turn, and whether anything can raise it.
- Combat state: health, block that fades when the turn ends, and status effects such as poison or weakness that tick down each turn.
- Enemy behavior: a move pattern for every enemy, with the next move on display.
- The run: a map generator, card rewards, a shop, rest sites and events.
- Persistence: what survives a run (unlocked cards, a best climb) and what does not.
Content
A first version needs less than you might expect: one hero, a starting deck of around ten plain cards, a draft pool of a few dozen, a dozen or so enemy types, a couple of elites, one boss and a small set of relics. That is one act. Make one good act before you plan three.
Tuning
These numbers decide whether the game is fun: starting health, energy per turn, cards drawn, the damage and block on basic cards, how quickly enemy damage climbs through the act, and how often rare cards show up as rewards. Balance in this genre is mostly about synergy. Two cards that are fine alone can break a run together, and hunting for those pairs is most of the testing.
Making your version in Tyrbo
Open the workbench at app.tyrbo.ai in your browser and describe the game. Name the hero, the resource, the enemies and the one rule that makes your deck different. Some starting points that already move away from the original:
- "A deckbuilder where a lighthouse keeper fights sea storms; cards are tides and lamps, and the deck gains a card each time the light goes out."
- "A roguelike deckbuilder in a night market, where every card costs coins instead of energy and you shop between every fight."
- "A gardener's deckbuilder: seed cards stay on the table and grow stronger each turn, while weeds the enemies plant clog your draw pile."
Tyrbo then asks about what your sentence left open, each question with a best guess already selected: portrait or landscape, how long a run lasts, how forgiving the first act is, what the cards look like. Keep the guesses that fit and change the rest.
Next comes the plan: a title, a pitch, every part of the game listed, and concept art. Read it as the designer. Do enemies show their intent? Are card removal and rest sites in it? Is a reward a choice of three, and can the player skip it? Anything missing is one reply away. Seeing the plan and concept art is free once you sign in; building runs on paid plans with monthly credits.
When you approve it, Tyrbo builds from tested parts. The hand, card faces and card play sit on its cards interface, health and energy on its HUD part, and each of your card effects is written as new code for your game. Scan the code on screen and the game runs on your phone.
Then tune it by sentence. Specific requests work best:
- "The first elite kills me every run. Cut its damage by a quarter."
- "Nobody picks the poison cards. Make poison last one turn longer."
- "Add a test panel that puts any card into my deck and any relic into my bag."
That last request matters most. A panel that adds cards on demand lets you test a combo in a single fight instead of hoping to draw into it over twenty runs. For more on the genre in Tyrbo, see the deckbuilder game maker and the wider card game maker.
Making it yours, not a copy
The bones of a roguelike deckbuilder (energy, a hand, a branching map, card rewards) are now shared by many games, and game mechanics are generally free to build on. This is not legal advice, but the usual line runs between an idea and its expression. What you must not take is Mega Crit's expression: the name Slay the Spire, its logo, its characters, its card names and card text, its relic names, its enemies, its art and its music. Do not set your game in a spire, do not reuse a well-known card with one number changed, and do not put their name in your title or store description to borrow attention.
Google Play's policies reject apps that impersonate other apps or infringe someone's intellectual property, so a game that leans on another game's name or art risks removal. Read the Google Play intellectual property policy before you list anything.
There is a design reason to go further than the legal floor, too. The deckbuilders that found their own players each changed a core rule: what the resource is, where cards go once played, how the map works. Try one:
- Change the resource. Health spent as mana, a pool shared with the enemy, or energy that carries over when you save it.
- Change where cards live. A board where played cards stay, a hand that never discards, or a deck whose top card you can always see.
- Change the climb. A map that shifts as you move, rooms that close behind you, or one long fight across several phases.
- Change the theme completely. A cooking deck, a courtroom deck, a band on tour. A strong theme suggests mechanics of its own.
Choose one and build the whole game around it. Your prompt should describe your idea, not someone else's game.
Publishing it
A Tyrbo game is checked before it leaves the pipeline: no soft-locks (in a card game, a turn that cannot end or a reward screen with no way out), steady play on low-end Android phones, ads and in-app purchases wired in, leaderboards live for the highest climb, and a signed build you own with full commercial rights. Android is the first store, and a step-by-step checklist comes with the build. The guide on how to publish a game on Google Play covers the listing, the content forms and testing.
Questions people ask
Can I legally make a game like Slay the Spire?
Game mechanics such as energy, card drafting and a branching map are generally free to build on, and many deckbuilders use them. What you cannot use is Mega Crit's expression: the name, logo, characters, card and relic names, card text, art and music. This is not legal advice; if you are unsure, ask a lawyer.
How long does it take to make a game like Slay the Spire?
It depends on how many cards, enemies and acts you plan, and on how much balancing you do. A small first version with one hero and one act is the sensible start. Most of the time goes into playing runs and adjusting card numbers, which is where a deckbuilder becomes good.
Do I need to code to make a deckbuilder?
No. You describe the hero, the cards and the run in plain words, review a plan with concept art, and Tyrbo writes the code. Changing a card cost or an enemy pattern is a sentence. You can also ask for a test panel that adds any card to your deck.
Can players fight each other in my deckbuilder?
No. Card games built with Tyrbo are single-player against computer opponents, with leaderboards to compare runs. Design your enemies and elites to carry the challenge.
Make the game this guide is about.
Make your gameSign in to get the plan and its concept art. Both are free.
Slay the Spire is a trademark of its owner. Tyrbo is not affiliated with or endorsed by them; the name is used only to describe and compare.