We're two people building an action game in Unity called KUTO: The Lock of Time. Denys writes the code, Serhii does level design, animation, and the trailer. That's the team.
The pitch is simple to say and annoying to build: time isn't a feature in the game, it's the thing you play with. Not "travel to the past and the plot reacts" — the abilities you press in combat, in jumps, in exploration are all manipulations of time. The moment we kept designing around: you blow a jump, you're about to die, and instead of a loading screen you rewind a few seconds and retry the gap with what you just learned. Death becomes part of the loop instead of a reset.
That sentence hides most of the engineering. So here's the honest version.
Starting on Corgi, and where it stopped fitting
When you're two people, you don't write a platformer core from scratch. We started on Corgi Engine — movement, physics, and basic combat out of the box. For the first stretch it did exactly its job: we could prototype instead of plumbing.
Then the thing the whole game exists for — the time mechanics and a heavier combat system — outgrew what Corgi gives you as-is. That's not a complaint about Corgi; it's a great way to start. It just wasn't the right foundation to carry the whole game.
What we did about it:
- Kept the combat bones (collision, hit resolution and the like) because they were solid.
- Adapted those to our own setup with 3D layers instead of a flat 2D plane, so the 2.5D space and our own movement could sit on top.
- Built the rest — the time systems — as our own thing, because they touch everything and couldn't live as a bolt-on.
If you've extended an asset-store framework past its intended scope, you know the exact point where "configure it" turns into "rewrite it." We hit that point and stopped pretending otherwise.
Why "rewind" is harder than it looks
The interesting problem with Recall (the rewind ability) isn't playing time backwards. It's deciding what counts as state worth remembering.
Rewind means every meaningful thing in the scene has to remember its recent past — position, velocity, animation, health, whatever the player would notice snapping back. The actual design work is drawing the line between "the player will feel it if this rewinds wrong" and "nobody will ever notice." Get the line wrong in one direction and rewinds feel cheap and inconsistent; wrong in the other and you're recording the world every frame for no reason.
That tradeoff — fidelity versus cost, and what the player's eye actually tracks — is most of what makes a time mechanic feel good versus janky. We're still tuning it.
Five Time Keys, two slots
The time mechanics are five abilities, each removing one rule of reality:
- Recall — wind time back a few seconds.
- Dilation — global slow-mo. Bullet-time for precise jumps and fast enemies.
- Leap — a long dash forward through time, into spots normal movement won't reach.
- Fracture — time breaks gravity. Walk on ceilings, flip the arena, open new routes.
- Stillness — stop time but keep moving yourself.
The design constraint we like most: you carry only two active Keys at a time, swappable only at your base. So you commit to two abilities before a run and live with that choice until you die. It keeps each Key meaningful instead of a crowded ability bar, and it gives the meta-progression something real to chew on.
Death that keeps some of your progress
It's a roguelite loop: run, die, restart. Death doesn't wipe everything. Across runs you keep your Keys, your meta currency, unlocked eras, and forged weapons. What resets is Key fuel, your basic weapon, and that run's random bonuses.
Balancing that split — keep enough that a death isn't wasted time, take enough that it stings — is the part we rewrite most often.
Where it's at
Early alpha. We're heads-down on one demo level, short but polished enough to teach all five Keys without a tutorial wall.
If the rewind-instead-of-reset idea sounds like your kind of thing, that's exactly the reaction I'm trying to find out about this early. You can wishlist it on Steam, and I'll keep writing these as the systems settle.
Happy to get into specifics in the comments — especially from anyone who's built rewind or recording systems and has opinions on where to draw the state line.
Top comments (0)