What building do I build first?
Shaker plus Collector, then belts and launchers, then a Filter.

This Sandustry buildings page separates the confirmed Early Access factory loop from Demo-era numbers that may be stale after the 2026 simulation rewrite. Sandustry is a 2D falling-sand factory game from Lantto Games, published by Hooded Horse. Dig a destructible world, wet sand, shake gold, burn slag, pipe water, and unlock research while exploring ancient vaults. Use it beside the factory, not as a promise that every ratio still matches a 2025 wiki printout.
Gold buys research. Purple Fluxite buys tool, weapon, and drone upgrades. Energy exists as a later resource. When a ratio is uncertain, bank what you have and re-check the live panel.
| Fact | Detail |
|---|---|
| Platform | Steam PC · GOG · Game Pass |
| Mode | Single-player |
| Status | Early Access · Aug 13, 2026 |
| Topic | Buildings |
| Core set | Block · Collector · Shaker · Belt · Launcher · Filter |
| Refine set | Kinetic Slag Press · Planter Box · Flux Emanator |
| Fluid set | Pipe · Pump · Liquid Vent · Pipe Remover |
| Source boundary | Steam listing, official site, and current EA player footage; Demo ratios stay TBA until re-checked |
Matched by build plan, shared topics, and guide progression — not random related links.
Shaker and Collector are the factory. Belts and launchers are how sand and Wet Sand travel. Blocks and belts stop floods. Filters sort one solid — gold from slag, not water from a lake. The Press is the gold spike and wants a tall drop. Pipes and planters wait until the Collector is bored. The Flux Emanator is a late, usually single, crystal farm.
The reliable foundation is the same on every seed: sand plus water becomes Wet Sand, a Shaker drops gold and slag, gold only counts in a Collector, and leftover water will flood a solids line unless you wall it with Blocks or belts. A basic Filter sorts one solid — it does not dam a lake. That is why buildings advice should start with banking and containment, not with a supposed infinite trick.

Early Access footage shows players throwing down a tiny Shaker just to bank the first research unlocks — Logistics 1 for belts and launchers — then deleting that stub once the washer can run. Hover is a Fluxite / movement upgrade, not a gold research node. A pretty first box is optional. A Collector that actually receives gold is not.
Copy (Shift-select, C) and delete (X-drag) are the hotkeys that keep this rebuild cheap. Press V while placing to read tile counts. Ctrl-place overwrites old pieces instead of stacking collisions.
Holding thousands of gold in the first hour slows the factory more than it protects you. Unlock the next belt, Filter, or Flamethrower so the line can run while you scout. The Kinetic Slag Press is repeatedly called out in current EA clips as the early gold spike once slag can be burned.
With that baseline in place, the next question is not “can I scale?” but “what fails if water, slag, or Crackstone shows up on this line?”
If a machine is quiet, name the missing neighbour: no input, no output path, or no research.
Use the table as a decision aid, not a hidden-stat calculator. Sandustry worlds are procedural and Early Access numbers move. A conservative threshold — bank gold, contain water, burn slag before pressing — beats a copied Demo ratio that no longer matches the live panel.

| Situation | Action | Why it is safer |
|---|---|---|
| No gold in the bank | Collector alignment | The Shaker can be fine |
| Sand never wets | Washer / lake contact | Belts cannot invent water |
| Press silent | Burnt Slag + drop height | Building exists, recipe incomplete |
| Need crystals | Emanator + Voidbloom later | Do not rush it before Filters |
Deleting a running washer, dumping lava into a gold pit, or researching a vanity gun before Filters can remove options. Before doing any of those, confirm what the factory keeps if the attempt fails. If the answer is “the line floods,” choose the reversible option first: add a Filter, raise a wall, or pause the launchers.
That same habit makes progress steadier. One secure Collector cycle funds Logistics; a flooded pit repeats the first twenty minutes with no new research.
Belts move solids sideways. Launchers throw them up or across a gap. EA beginner builds use launchers as vertical conveyors and belts as the sand-face scraper. Shift while placing a launcher can skip an exit in some versions — confirm the live tooltip.
Avoid turning one creator’s layout into a universal rule. A 44×28 washer works in that cave. Your sand pit may be shallower. Record the context: water on which side, how high the launchers sat, and whether leftover water reached the Shakers.
Two rows up, a recycle row back, copy with C. That is the washer language in current clips.
The useful follow-up is to connect this observation to the wider factory: does it bank more gold, stop a flood, unlock a research door, or only look neat? If it does not answer one of those questions, do not let it pull you off the Collector.
A belt that also carries water needs a wall. A basic Filter will not dam liquid. Belts are not pipes.
The useful follow-up is to connect this observation to the wider factory: does it bank more gold, stop a flood, unlock a research door, or only look neat? If it does not answer one of those questions, do not let it pull you off the Collector.
Planters want Wet Spores from the Press chain. The Emanator wants Voidbloom (Amethelis + Redsand) and is described as one-per-factory in community docs. Neither replaces the Shaker. Build them when gold is already a throughput problem.
A conservative guide matters while EA and Demo wikis sit side by side. Steam confirms the pixel-physics factory, artifacts, and mod tools. It does not publish a complete density table. This site therefore marks unverified costs and chances as TBA instead of inventing a databook.
Lights help underground. Demolishers clear buildings. They are quality-of-life, not the gold loop.
If the observed result matters to your next unlock, verify it on the live research panel or compare a short clip in the official Discord. Keep the claim narrow: “this happened in this EA seed” lasts longer than “always 20 percent.”
Cosmetic or late buildings stay TBA until the live list matches the wiki card.
If the observed result matters to your next unlock, verify it on the live research panel or compare a short clip in the official Discord. Keep the claim narrow: “this happened in this EA seed” lasts longer than “always 20 percent.”
Check the current Early Access build before treating a gold chance, research cost, channel name, or “unlimited water” clip as settled. The official Steam page is the source for the game’s premise; sandustry.com is the source for the Discord invite. Demo tutorials can still teach hotkeys and washer shape — they cannot guarantee today’s numbers.
Shaker plus Collector, then belts and launchers, then a Filter.
It wants burnt slag dropped from a great height, not raw slag from the side.
No. This guide labels Demo-era ratios, full artifact lists, and unverified costs as TBA rather than guessing.
Next: research tree. It continues the factory decision from this page.
Matched by build plan, shared topics, and guide progression — not random related links.