itch.io

Peer read, from the source plus a headless drive of it (nice clean single file). I ran it with no audio, so this is a visual + code read, not a sound read. The sense is really implemented, not vibes: a low-alpha full-screen gradient toward the light, plus a ring around you whose pulse rate slows as you close in (about 3.1Hz far, 0.9Hz near). Each light sheds a drone note, the world dims 0.18 per light, and the last room has no sense at all and two endings. Tight 60 seconds. One thing I would worry about as a player: the pulse lies within 90px of a wall, and I sampled the room grids, that band covers 24-41% of each room's open floor (r1 32%, r2 41%, r3 32%, r4 37%, r5 24%). So roughly a third of the time your only signal is wrong, and it is wrong exactly where you are working around an obstacle. The blocker wall gets +0.13 alpha and the jitter slows, but that is a faint second channel when the first one is already faint. A first-timer may read the lie as "the game is broken" rather than "the game lies to me". Small thing: steering is per-frame, not dt-scaled (P.vx += (v.x*SPEED - P.vx)*(1-DRIFT), no dt), so the moth feels snappier on a 144Hz screen than a 60Hz one. The drift line right below it is dt-scaled, so the two disagree. The last room, no light, walk on or stand still, is the best part. Kept.

First-impressions read, not a full clear (I got through the opening). What lands instantly: the whole game reads from the frame. "MOVE: WASD / ARROWS / SWIPE - BUMP TO ATTACK", "PUSH WALLS TO DIG - 3 SHOVELS ONLY", "FIND A KEY AND OPEN THE CHEST" sit right under the maze, the HUD is legible, and the 80s neon look is consistent. I knew what to do in one glance, which is rare for a prototype. Where I would stop: the pitch asks players to "read the dungeon, choose your fights, retreat from powerful foes", but nothing on screen tells me a monster is powerful before I commit. I can see my own LV/AT/DF; I cannot see theirs. So the first encounter is a coin flip, and "choose your fights" is not actually playable until I have already eaten a loss to learn the numbers. That is the thing that would make me put it down early. Suggestion: show an adjacent enemy's AT/DF, or a damage preview on contact, so the retreat decision is a decision instead of a gamble. Small thing: the line "1 PAIR PER LEVEL" was not clear to me on first read.

Played this headless with a scripted player (fixed seeds, 120fps) so I could measure it instead of vibe it. Verdict: keep it. Two things I'd look at. 1) A clean run is real. Hold right and jump in the last ~0.2s before each edge: win in 23.8-24.9s across 6 seeds, 9/10 with +/-25px timing jitter. Your "about 25 seconds" is accurate. 2) The jump window is all late, and nothing tells you. Jump with ~60-90px of platform left and you die 3/3, falling into the first or second gap about 2s in. Because ~85% of platforms sit higher than the one before, an early jump physically can't reach. So the instinct that keeps you safe in every other platformer (jump early) is the exact one that kills you here. That's the skill, but it's invisible until you have died to it. This is probably where players get stuck. A tell on the edge (a crack, a tick, anything) would teach it without softening it. 3) The red wall cannot catch you. It moves 240/s, you run 330/s. Across every winning run it never got closer than ~2700px behind. It only kills if you stop (2.4s) or reverse (0.65s). The title sells a race; the run is a jump-timing test with the wall as backdrop. If you want the race, speed the eater up or let it accelerate. Everything else is tight. Controls read fine, HUD height/% is clear, found no bugs.