Mycopunk

Mycopunk, still in Early Access, has a problem most games would envy: its best idea is so good it exposes everything around it.

The setup is standard-issue Devolver Digital irreverence. You are one of four disposable robots, contracted by a corporation called SAXON, sent to an alien planet to exterminate a fungal infestation that is very obviously going to outlive you. Nobody in this fiction respects you. You are equipment with a personality. The writing knows this joke and mines it hard — the comedic register is closer to workplace-humiliation comedy than to the chest-thumping military parody that co-op shooters usually default to.

Structurally, it’s a mission-based co-op looter shooter for up to four players, with procedurally arranged contracts, an orbital dropship hub between runs, and a solo option that technically exists. The genre lineage is unmistakable: Deep Rock Galactic set this template, Helldivers 2 scaled it, and every extraction-flavored PvE shooter since has been arguing over the leftovers. Pigeons at Play — a small studio with a name that suggests they know exactly how seriously to take themselves — did not try to out-scale anyone. They went after the part of the genre nobody had fixed: progression.

Because here is the dirty secret of the co-op shooter. The shooting is usually great and the character-building is usually a menu. You unlock nodes. You pick a percentage. You close the tab. Mycopunk’s answer is a grid-based upgrade system where you physically place, rotate and overlap hundreds of hand-crafted modifiers onto your weapons and gear like you’re packing a suitcase for a war crime. The game ships with a dozen weapons, a few grenade types and hundreds of upgrades, and the upgrades are the number that matters — not because the count is large, but because they’re authored rather than generated, and because fitting them together is itself the game.

Steam sentiment has been positive overall, with recent readings softer than the lifetime average. That modest footprint tells you something honest — this is a cult hit, not a phenomenon, and the 1.0 launch will decide whether it stops being one.

• The Grid Upgrade System: A spatial inventory board per weapon and gear slot. Modifiers are irregular shapes. You rotate them, wedge them, and — critically — overlap them, which is where the system stops being Tetris and starts being a puzzle with a damage number attached. Chaining overlaps into compounding effects is how you build something broken.
• Four Asymmetric Robots: Each character carries a distinct traversal identity — grapples, jetpacks, drills — that reshapes how you read a map, not just how you fight in one. Movement is the class fantasy here, and it’s the right call.
• Procedurally Arranged Contracts: Missions assemble from authored components rather than pure generation, keeping encounters legible. Up to four players drop in; solo is supported, with caveats.
• The Orbital Dropship: A downtime hub for loadout tinkering and unstructured squad goofing. Low-stakes, high-charm, and structurally more important than it looks.
• Early Access Pricing at $14.99: Priced like a demo, stocked like a full game’s worth of systems.

The grid is the whole argument

The loadout screen alone justifies the purchase for a certain kind of player.

Most looter shooters express power as a scalar. You had 100 damage, now you have 118 damage, congratulations. Mycopunk expresses power as geometry. A modifier isn’t a stat line — it’s a shape with a footprint, and every shape you commit to costs you space you might have wanted for something else. The overlap rule turns that constraint inside out. Instead of purely subtracting from your options, careful stacking multiplies them, so the optimization problem has a genuine skill ceiling. Bad players fill the grid. Good players interlock it.

This is the most interesting take on progression the genre has seen in years, and the reason is architectural. The build screen has become a gameplay mode. Deep Rock’s overclocks are excellent but they’re still a selection problem. Path of Exile’s tree is vast but it’s still pathing. Fitting irregular polygons into a bounded board is a fundamentally different cognitive activity — tactile, visual, immediately legible — and it hands you the “I broke the game” moment that these games live on, without requiring a spreadsheet to find it.

The upgrades are hand-crafted, and it shows in a way procedural affixes never do. These aren’t +4% variants padding a loot table. They’re effects with personality, and enough of them interact that discovering a combination feels like discovery rather than arithmetic.

The gameplay loop, and where it thins

The moment-to-moment shooting design earns its keep. The guns are built with weight and snap. Traversal is fast and expressive — the jetpack and grapple characters play like they escaped from a movement shooter and wandered into a horde game, which keeps firefights mobile instead of turret-y. Fungal enemies are designed to pop and spread in a satisfying, faintly disgusting way. Nothing here is stiff on paper.

The trouble arrives on a schedule. Past the early stretch of contracts, the mission variety stops surprising. Objectives begin to reveal their small vocabulary. The procedural arrangement reshuffles space without reshuffling intent, so the twelfth contract asks the same question as the fourth in a different room. That’s a content problem, not a design failure, and the distinction matters — but it’s still the wall, and it arrives right as grid mastery peaks. You finally know how to build something ridiculous, and the game runs short of places worth taking it.

The softer recent Steam sentiment reads exactly like that wall. It isn’t a game getting worse. It’s a game whose long-tail players exhaust it faster than updates arrive.

Solo play is the weakest seam

The difficulty curve does not scale down cleanly. Encounters are tuned around the crossfire and revive economy of a squad, and stripping three robots out of the equation doesn’t just make fights harder — it makes them different, in ways the balance hasn’t accounted for. Some contracts become attrition slogs. The asymmetric class design, which is a strength in a group, becomes a liability alone: you’re locked into one traversal answer for problems that were authored with four.

If you’re buying this to play by yourself, adjust expectations. It works. It isn’t what the game is for.

Matchmaking is the unglamorous blocker

Lobby stability and matchmaking are, by consistent player account, rough. Drops, failed joins, sessions that don’t hold. For a game whose entire value proposition is four people goofing off in a dropship, friction at the front door is the most expensive kind of bug there is. No amount of build-crafting brilliance survives three failed connection attempts. This is the single most important thing on the pre-1.0 list, and it’s more urgent than any content drop.

The dropship earns its place

Between-mission hubs are usually a menu with furniture. Mycopunk’s is closer to a green room — a space designed to let four people be idiots together while someone rearranges their grid. It’s cheap to build and disproportionately effective at making a session feel like hanging out rather than queuing. More PvE shooters should steal this.

The art direction is grimy retro-futurism done with actual conviction: chunky industrial silhouettes, muted rust-and-mustard palettes, machinery that looks maintained by the lowest bidder. It’s specific, and specificity is rare. The robots carry the whole tone — expressive, over-animated in the right way, with squash-and-stretch personality doing work that the (very funny) writing then doesn’t have to do alone.

The stylized rendering rather than fidelity-chasing is a sensible choice for a game of this scope, and the stability complaints in the player base cluster around networking, not framerate — a meaningful distinction, since one is a tuning problem and the other is a rewrite.

The UI deserves its own note. The grid screen is built to stay legible under pressure, communicating fit and overlap without a tooltip archaeology dig, and making a complex system feel physical. That’s hard. Most studios ship worse inventory UI in games that cost four times as much.

Steam is currently the only home, and the game is built like it knows that. The grid system is a mouse-native interface — drag, rotate, nudge, compare, reconsider. A cursor is the correct instrument for it, and the design has clearly been iterated against one.

Controller works for combat, where the fast traversal and generous target profiles are forgiving. It works less well for the loadout board, where precision placement across a dense grid becomes a stick-driven negotiation. Any future console port will live or die on this: the studio’s signature system is the one that ports worst. A controller-first redesign of the grid — snapping, shape-cycling, radial placement, something — isn’t a nice-to-have. It’s the whole port.

Mycopunk is the rare Early Access title where the ambitious part is done and the boring part isn’t. That’s backwards from the usual failure mode, and it’s the reason for measured optimism. Systems are hard; content is expensive but tractable. Pigeons at Play have already solved the problem that requires actual invention — they built a progression system other studios will be copying within two years — and what remains is the grind of authoring more missions, fixing netcode, and tuning a solo curve that currently doesn’t work.

The honest read: this is a solid game with roughly half the content it needs, sold at a fraction of the price it could eventually command. We rate it a Good — recommend it to squads with patience, and to everyone else once 1.0 lands. Buy it now if you have friends and patience. Wait if you don’t.

Either way, watch the grid. Someone is going to steal it.