Why Inscryption Bet on Unity: The Boring File-System Tech Behind Card Horror's Nastiest Twist
Short answer: Inscryption runs on Unity, listed as the Engine in its infobox alongside developer Daniel Mullins Games and publisher Devolver Digital. The game was directed by Daniel Mullins and originally released for Windows on October 19, 2021, then ported to Linux, macOS, PlayStation 4, PlayStation 5, Nintendo Switch, and Xbox One and Series X/S over the next two years.
That choice looks lazy in 2026. Every serious card-game postmortem says to pick a lean 2D framework and polish one loop forever. Mullins did the opposite: he used Unity to ship three half-finished card games, a first-person escape room, a pixel-art RPG, and a found-footage ARG in one $20 package — then sold 1 million copies by January 2022. The lesson is not that Unity is good at cards. It is that Unity is good at pretending to be three different engines while quietly opening your hard drive as a second game board.
From jam card to three economies
Inscryption started as Sacrifices Must Be Made, a much shorter deck-builder Mullins made for a Ludum Dare game jam, where you had to kill your own creatures to summon others. After publishing the prototype to itch.io in December 2018, the positive response led him to expand it into Inscryption.
Ludum Dare is a game-jam competition where participants must create a video game fitting a theme in two or three days, with all code and content created during the competition by a single person. That origin explains the core economy: one 3x4 grid later expanded to 3x5, attack and health on each card, unblocked damage tipping a weighing scale by five teeth.
What Mullins added with Unity was not more cards, but more payment methods:
- Act 1 (Leshy's cabin): blood sacrifice from cards already in play, plus getting up from the table to solve escape-room puzzles and find the card forms of P03, Grimora and Magnificus. Lose twice on a map or once on a boss and Leshy takes you with his camera and turns you into a death card that can appear on later runs.
- Act 2 (pixel overworld): presented as a pixel-art top-down RPG similar to Pokémon, where you pick one of four Scrybes to replace, collect packs, and challenge all four Scrybes. Here cards from all four Scrybes are available, and losing a card game carries no penalty.
- Act 3 (P03's factory): a roguelike deckbuilder again, but now against P03 in a robot factory trying to achieve Transcendence by publishing Inscryption on the Internet. P03 can use death cards you created in Act 1, and losing resets you to the last antenna-tower checkpoint.
The rulebook phrases it dryly: those created by P03 use energy with refilling cells, those by Leshy require blood sacrifice, those from Grimora require bone tokens earned when cards are defeated, and those from Magnificus require gems on board to be played. In other words, four resource systems for the same 3-column combat.
Most reviews treat that as bloat — and Act 2 to Act 3 is where criticism concentrates. My read is different: the bloat is the port.
What Unity actually bought
Unity is a cross-platform engine developed by Unity Technologies, first released in June 2005 for Mac OS X, particularly popular for iOS and Android mobile development, considered easy for beginners and popular for indie development. It handles 2D and 3D games and interactive simulations.
Mullins did not need Unity for card physics. He needed three boring Unity features:
1. Cheap scene swaps as genre swaps. Act 1 is first-person in a cabin, Act 2 is top-down pixel art, Act 3 is grid rooms like The Binding of Isaac. In a custom C++ engine that is three renderers. In Unity it is three scenes with different cameras and the same C# card logic underneath. That is why the game can change perspective from first-person to third-person to 2.5D without rewriting combat.
2. System.IO as horror. The game is presented as found footage recorded by Luke Carder, The Lucky Carder, who finds Inscryption on old out-of-print cards plus handwritten coordinates. Its presentation mixes computer animation, pixel art, drawn art and live action, with no backstory at start and a metafictional narrative plus a real-world alternate reality game embedded both in video and outside it. None of that works if you cannot read and write files, fake error screens, and pretend the game is haunted. Mullins already proved the trick on Pony Island, a 2016 metafictional game also built on Unity, where Lucifer traps you in an arcade cabinet, Hopeless Soul talks to you via instant messenger, asks you to delete three core files, and at the end talks directly to you and asks you to delete the game file. Pony Island's prototype was made during a Ludum Dare jam and later passed Steam Greenlight for a January 4, 2016 release. Inscryption is Pony Island with cards: Unity's file access is the monster.
3. One store page, eight ports. Daniel Mullins Games lists Inscryption as AVAILABLE NOW on Steam, GOG, Epic, Xbox, PlayStation and Switch from one hub page. Unity did not make the cards better; it made the same haunted cabin shippable from Windows to Switch (December 1, 2022) to Xbox (April 10, 2023) without a rewrite. For a solo dev, that is hiring, QA and console certification in a box.
My contrarian take: impurity as a moat
The standard advice — echoed by every Slay the Spire clone postmortem — is that Mullins should have shipped Kaycee's Mod first. Compare the discipline:
Slay the Spire is a 2019 roguelike deck-builder by two-person Mega Crit, built on libGDX, released January 23, 2019 after early access in 2017. You pick one of four characters, climb branching floors to a boss, play attacks for energy, and dilute or trim your deck on the fly. It is considered the game that popularized roguelike deck-builders. One loop, perfected.
Inscryption deliberately ships impurity: a roguelike that quits being a roguelike in Act 2, a horror game with no fail state in the middle, a card game that lets you stand up and leave the table. Conventional wisdom calls that pacing suicide.
I think it is Unity-specific camouflage, and I will state it bluntly even though it is not provable from telemetry:
- Sacrifice is garbage collection. Blood, bones, energy and gems are not four designs. They are one teeth-scale with different GC pressure. Blood forces you to delete your own board, bones reward you for things already dying, energy rations you by turn. Unity cannot simulate 18 quintillion planets, so Mullins makes you the memory manager. You feel clever when you are just freeing RAM.
- The cabin is a loading screen you can walk in. Between encounters you solve cabin puzzles like an escape room. That looks like extra content. Technically it is a corridor that hides deck reshuffles and map generation while keeping you in first person. Other Unity horror games hide loads behind doors; Mullins hides them behind a safe combination.
- Three bad engines beat one good engine. A libGDX purist would have to justify Act 2's no-penalty RPG. A Unity generalist can shrug: scenes are cheap, so ship the prototype. Players remember the switch, not the balance. The free Kaycee's Mod expansion in March 2022, letting players focus on the tabletop without the other elements, proves the point backwards — once the trick was revealed, players demanded the boring perfect loop Mullins had refused to sell them.
In short: Slay the Spire used a boring framework to make cards sacred. Inscryption used a boring engine to make cards suspicious. That is a different job, and only Unity's VCR qualities — swap tapes, fake tracking errors, record over your save — make it affordable for one person.
What this cost, and why it still worked
Gains:
- Jam-to-Steam pipeline in one language. The Ludum Dare rule — single person, 48 hours, own the game — scales directly to Unity C# solo dev without hiring engine staff.
- Horror without lighting. Low-poly cabin plus film grain plus live-action vlog hides Unity's lack of AAA ray tracing. Darkness is cheaper than Nanite.
- ARG for free. Real-world coordinates, decrypted easter eggs and fake extraneous files cost forum threads, not engine features.
Costs:
- Balance whiplash. Act 1's permadeath teeth race does not teach Act 2's penalty-free collection or Act 3's checkpoint grind. New players bounce at each switch.
- Porting debt. Eight SKUs from Windows to PS5 to Switch mean Unity version pinning and console cert for a game that pretends to corrupt itself — a QA nightmare if your horror is file corruption.
- Sequel trap. You cannot make Inscryption 2 by adding cards. You have to invent a new lie about what the software is. Pony Island to The Hex to Inscryption shows Mullins knows this; each game lies about being a different kind of software.
What indie devs can steal
- Prototype the payment, not the card. Before art, test blood vs energy vs bones on paper. If sacrificing your best Squirrel does not hurt, no sigil will save you.
- Use Unity scenes as chapters, not levels. One combat class, three presenters (cabin, overworld, factory) is cheaper than one perfect board.
- Make failure content. Leshy's camera that mints your failed run into a future death card turns rage-quits into collection. Steal that before you steal sigils.
- Ship the ARG in the patch notes. You do not need a real-world conspiracy. One coordinate in a trailer plus one deletable file is enough for Reddit to do the rest.
- Plan the pure mode on day one. Kaycee's Mod was not DLC; it was an apology to roguelike purists. If your genre-switch will annoy half your players, build the toggle before launch.
Try the loop in your browser
You do not need to install anything to feel why sacrifice works. These instant-play trainers isolate each skill Inscryption demands:
- If the blood economy confuses you, grind patience in Classic Card Games Collection, a 150-plus solitaire collection with Klondike, Spider and Freecell. It teaches the same lesson Leshy teaches: winning means knowing what to throw away, not what to keep.
- If sigils blur together, slow down with Card Shuffle Sort, a calming color-sorting puzzle with daily challenges. Sorting by color under rising difficulty is Act 2 deck-trimming without the jump-scare — organization as damage.
- If you keep ignoring the cabin, practice standing up in Rooms Home Escape, where you wake in a locked house and must link a kitchen note to a living-room pattern to escape. That drawer-to-painting logic is exactly how you find P03, Grimora and Magnificus between card battles.
Play one round of each, then return to Leshy's table. You will sacrifice faster, read the scale earlier, and — my unrigorous but repeatable observation — you will stop blaming the RNG and start blaming your own hoarding, which is precisely what Mullins wanted Unity to make you feel.
Top comments (0)