ASTLIBRA Revision

Judging games by their storefront thumbnails is a bad habit, but an honest one. Most of us do it. And by that metric, ASTLIBRA Revision should be dead on arrival: stock-looking sprites, backgrounds that appear to have been pulled from a fifteen-year-old asset pack, a UI that seems allergic to whitespace. It reads as a weekend project someone shipped for pocket change.

It is, in fact, the opposite. ASTLIBRA is the product of a many-year solo development effort by Japanese developer KEIZO, plus further remastering work before WhisperGames brought this Revision to a global audience on October 13, 2022. That’s not marketing trivia — it’s the single most important fact about how this thing plays. Games built by committees converge toward legibility and safety. Games built by one person over many years converge toward obsession, and ASTLIBRA is obsession made playable: an action RPG that keeps stacking systems on top of systems long past the point where a publisher’s producer would have said “enough.”

The pitch is deceptively plain. You’re a young man ripped out of his home and dropped into a hostile, unfamiliar world, with a talking bird named Karon as your guide, conscience, and exposition delivery system. From there it’s a chapter-structured march through dungeons, ruins, and boss arenas, hunting for a way back to a life that may not exist anymore.

Market positioning is where it gets interesting. This sits in the same shelf space as Momodora, Salt and Sanctuary, and the cheaper end of the Metroidvania aisle, but it isn’t really competing with any of them. Those games sell you tight, curated design. ASTLIBRA sells you volume — an absurd amount of it, at a $24.99 price that makes the comparison feel unfair. Its storefront reception has been strongly positive, and the pattern is consistent: players who bounce off it bounce in the first thirty minutes. Players who don’t tend to stay for the long haul.

• Time, memory, and fate as mechanics, not metaphor: The story’s obsessions are wired directly into the systems. Progression, loss, and recurrence are narrated at you between fights, and they also show up in what you unlock, what you keep, and what the game takes back.
• Tactile three-verb combat with Possession Skills: Slash, thrust, and smash form the melee grammar, and enemy waves feed a meter that unlocks Possession Skills tied to the foes you’ve killed. The combat is designed to be fast, weighty, and readable — the best-tuned thing in the package.
• A stacked progression architecture: An experience grid, force points, forgeable and upgradeable equipment, item synthesis, and collectible gear all layer on top of each other. Builds genuinely diverge into melee brawler, ranged specialist, or staff-wielding caster.
• Six difficulty tiers: From casual to punishing, adjustable to fit how much grind you’re willing to absorb.
• The Postscript: A substantial post-credits campaign with new story, new systems, and new challenges. Not a boss rush. An actual second act.

The gameplay loop

Here is the trick ASTLIBRA pulls, and it is a smart piece of design psychology. Nearly every action in this game moves at least two bars forward at once.

You clear a room of enemies. That fills your Possession meter, drops materials, ticks the experience grid, feeds force points, and pushes you closer to the next forge upgrade. Then you spend those resources, and the numbers on your weapon go up, and the next room dies faster, which fills the bars faster. The loop is nakedly Skinner-boxed and it works with embarrassing efficiency. Most RPGs make you choose between the fun thing and the optimal thing. ASTLIBRA collapses that distinction — killing things is the optimization path.

What keeps it from being hollow is that the combat underneath is built on solid foundations. The three-attack system is simple to describe and surprisingly deep to execute, because enemy spacing and animation commitment matter constantly. There’s real weight to a smash landing. There’s real punishment for whiffing one. The Possession Skills system, where defeating enemies grants you access to their capabilities, ties the grind to the moment-to-moment fantasy in a way that feels earned rather than transactional.

Where it demands patience

The opening chapter is bad. Not broken — bad in the sense that it front-loads every one of the game’s weaknesses and none of its strengths. Presentation is at its most stock, the writing is at its most stilted, systems haven’t opened up yet, and your character hits like a wet towel. The onboarding friction is severe, and it is unquestionably why some portion of players refund it. If the community consensus says “give it two hours,” believe it, because the game does not begin to argue for itself until the second and third chapters.

The English localization is the other persistent scrape. It’s functional and comprehensible, but it’s uneven — occasionally clumsy, occasionally flatly awkward, and it undersells a story that is more ambitious than it has any right to be. When ASTLIBRA lands its emotional beats, and it lands several hard ones, it does so despite the prose rather than through it. That’s a real loss. A story about time and consequence deserved a translator with room to breathe.

Grind is the third complaint, and here we’ll defend the game more than most. Yes, there are difficulty spikes, and yes, there are stretches where the correct answer is “go get stronger.” But the six difficulty settings exist precisely to let you dial that out, and the encounter design usually telegraphs when you’ve walked into something you’re underleveled for. The spikes that frustrate are typically loadout problems, not stat problems — bosses in this game are pattern-heavy set pieces that punish the wrong build far more than they punish the wrong level.

Interface and information design

The menus are the weakest engineering in the package. Navigating the forge, synthesis, and equipment layers involves more cursor travel and more mental bookkeeping than it should, and critical comparison information isn’t always where your eyes expect it. When your entire hook is a progression system with five interlocking currencies, the menu is the game — and this menu was clearly built by someone who had memorized it years before you arrived. It’s learnable. It’s never elegant.

Structure and pacing

The chapter format is the smartest structural choice here. It carves a very long game into digestible arcs with clear objectives and clear endings, which does enormous work in keeping a project this size from feeling shapeless. Dungeons are long and stuffed with secrets, rewarding the kind of player who checks every wall. And the Postscript deserves specific praise: post-game content is usually a difficulty-scaled victory lap, but this one adds story, mechanics, and challenge substantial enough to function as a sequel that shipped in the same box.

Let’s be blunt about the art. ASTLIBRA looks dated, and a meaningful chunk of it reads as stock or RPG Maker-adjacent asset work. Character sprites, environment tiles, and effects don’t share a consistent visual language, and the result is a game that never coheres into a look you’d call designed. If visual polish is a hard requirement for you, this is a wall you won’t get over.

But there’s a distinction worth drawing between ugly and unreadable, and ASTLIBRA is only the former. Combat legibility — hitboxes, telegraphs, enemy tells, damage feedback — is consistently clear, which is the visual property that actually determines whether an action game feels fair. The boss designs, in particular, punch far above the rest of the presentation. Some of those god-like encounters are imposing, built with a scale and silhouette sense that the overworld art never approaches.

Technically, there’s little to worry about. This is a 2D game of modest scope, and it’s light on hardware even through dense enemy waves and effect-heavy boss phases. The soundtrack is a quiet standout — moodier and more melodically ambitious than the visuals prepare you for.

Steam is the definitive version, and it isn’t close. Higher resolution, full remapping, and — critically — the freedom to run it windowed on a big display where the dense menus have room to be read. Use a controller. The three-attack combat grammar was built for face buttons and analog input, and keyboard-and-mouse is the weaker way to experience it. The one caveat: low-resolution source art scaled to a modern panel gets soft, and ultrawide setups won’t find much love here.

Nintendo Switch is the more complicated recommendation. The chapter structure is a natural fit for handheld sessions, and the pick-up-and-play cadence suits commute-sized chunks. The problem is the UI. Menus that are merely cluttered on a monitor become taxing on a small screen, and the small text in the forge and synthesis interfaces is the kind of thing that wears thin the longer a session runs. Docked play fixes most of it. If you intend to play primarily handheld, go in knowing the interface will be your main antagonist.

ASTLIBRA Revision is a deeply unfashionable game, and that’s most of what makes it valuable. Nothing here was smoothed by focus testing. The art wasn’t modernized, the menus weren’t simplified, the grind wasn’t tuned down to a mass-market curve — KEIZO simply kept building the game he wanted for years and shipped it with the rough parts still attached.

The result is lopsided in a way modern releases almost never are. The presentation is the worst thing about it and the systems are among the best things in the genre, and those two facts sit in the same package with no attempt at reconciliation. You have to decide which one you care about more.

We rate it a solid recommendation for the right audience and nothing more: if you can accept a game that refuses to seduce you on first sight, ASTLIBRA repays that patience across a long campaign and then hands you the Postscript and asks for more. There’s an entire category of players who will never get past the store page, and their hesitation is understandable. They’re also missing one of the more generous action RPGs of recent years — flawed, rough, and worth their time anyway.