General Overview and Tips

Your first Shaker may look like it has joined a tiny labor dispute: Wet Sand piles up, Gold stops falling, and the Water supply has turned into a damp hill nearby. Start with a jam check. Stop the feed, clear both exits, and place a small clean batch directly on top of the Shaker. If that bare test fails, save, reload, and try again; v0.5.7 fixed several causes of stalled buildings, but some errors may still linger.

That repair captures the main idea behind Sandustry. Recipes alone are not enough. Sand, Water, heat, gravity, and each material's physical state decide whether a factory runs or quietly builds a monument to poor drainage. The goal is to make every contact point and output path easy to inspect.

Version label: v0.5.7 (Update #4). Use the machine behavior and Shaker chance below for this build, then retest ratios, pipe connections, and old layouts after an update. Seed label: none; no world route is used here. From the starting area through the first portal, the game uses randomized tiles joined by fixed connections. Areas outside those tile regions are fully procedural, so exact terrain and resource positions can differ between worlds.

The First Progression Loop

Your opening loop is Sand → Wet Sand → Gold → Research. When one Sand pixel touches one Water pixel, both are replaced by two Wet Sand pixels. Place Wet Sand on top of a Shaker. Each processed pixel becomes Residue and has a 25% chance to produce one Gold. Gold passes through the bottom, while Residue stays above and is pushed very slowly toward the Shaker's facing side.

Contain Sand and Water separately. Let them meet only in a small mixing chamber. Contact changes both pixels into Wet Sand, so an exposed Sand pile can consume a reservoir one touch at a time.
Feed Wet Sand onto the Shaker's top. The material must reach the working surface. Dry Sand or another solid can cover that surface without being processed.
Keep both exits clear. Leave collection space below for Gold and an open route across the top for Residue. The Gold roll happens once per Wet Sand pixel, so a short dry spell can occur even when the machine is working.
Catch Gold in Collectors. Gold only counts as spendable currency while it is inside Collector blocks. A glittering pile beside the Collector is pretty, but it will not buy research.
Spend Gold on the bottleneck you can see. Buy transport when material movement is slow, sorting when streams mix, or fluid tools when Water control is the problem. A new building will not clear an old blockage.

Build With the Physics

Gravity sets the basic route. Sand and Gold fall, Water flows toward low ground, and Steam rises. Put falling inputs above their process, liquid basins below likely leaks, and a clear overhead path wherever Steam must travel.
Contact starts specific reactions. Sand and Water must touch to make Wet Sand, and Wet Sand must rest on top of a Shaker to be processed. Thin feeds make the active layer visible and keep the wrong material from hiding underneath a large pile.
Heat and phase changes alter transport. Fire melts Ice into Water and changes Water into rising Steam. Water can also combine with Snow to form Ice, while Steam that reaches the sky condenses into rain. Plan a low basin for Water, headroom for Steam, and a landing area for rain or newly moved liquid.
Density separates some mixtures. Wet Sand has a higher density than Residue. On a downward staircase of Shakers, Wet Sand tends to work beneath the lighter Residue and settle against the machines while Residue rises toward the outlet. Test a short section before copying the layout; pile behavior and balance can change during Early Access.
Every phase needs space. A chamber that makes Steam needs an overhead exit, and a chamber that receives Water needs a sealed floor and enough capacity. A perfect reaction chamber with no output room is merely an expensive jar.

Four Early Priorities

Containment: Seal corners, line reservoirs, and keep loose Sand away from stored Water. Water can slip through unsealed corner gaps, and any Sand it touches will turn into Wet Sand.
Output clearance: Inspect the top, bottom, and facing side of a Shaker. Residue needs room above and to the side, while Gold needs room below.
Buffers: Add a small holding chamber before each major process. A buffer absorbs uneven flow and shows whether a shortage begins upstream or a jam begins downstream.
Overflow routes: Give every by-product a safe destination. A catch chamber is better than letting Residue back up across the Shaker faces and block fresh Wet Sand.

If a Machine Appears Idle

No input on the working surface: Trace the line backward to the last moving material. Check the buffer, transport direction, and mixing chamber.
Correct material nearby, but no processing: Remove anything between the material and the machine. “Almost touching” is factory poetry, not production.
Input piled high: Clear every output before adding more machines. For a Shaker, inspect the Residue route above and the Gold pocket below.
Water keeps disappearing: Look for exposed Sand, diagonal corner gaps, and low pockets outside the intended basin. Seal the reservoir, then test it with a small amount.
Mixed or trapped material: Stop the upstream feed, remove the wrong pixels, and hand-feed one clean batch from the correct side.
The bare machine still does nothing: Save and reload, then repeat the clean-batch test. Reconnect one transport segment at a time after the machine works. If it remains stalled in v0.5.7, preserve the save and use the in-game bug report.

Do not debug five machines as one mystery. Prove one conversion, then attach its input, output, buffer, and overflow in that order. The smaller test gives the sand fewer places to practice mischief.

Compact Factory or Giant Sand Laboratory?

For a compact factory: Build short vertical modules with visible input and output pockets. Use gravity for simple movement, keep materials in narrow lanes, and leave one cleanup gap beside each machine. Compact lines are easy to copy, but they need strict overflow control because one full pocket can block the whole stack.

For a large experimental build: Spread processes apart and contain each one as a separate test cell. Add larger buffers, spill basins, shutoff points, and empty space between hot, wet, and granular systems. Scale by copying a module only after both of its outputs clear. The extra space makes it easier to stop one feed before a spill crosses into the next process.

Neither style is the one correct way to play. A tidy processor and a glorious sand cathedral can both work. The winning habit is the same: make each physical step visible, give every material a destination, and retest the module whenever a patch changes the rules.