Roadmap
One thing here is a promise. Everything past it is a plan, and we’re telling you which is which rather than letting a nicely formatted table blur the difference.
The one committed gate
G1 — “The Expedition” (milestone M2a) — playable, week 30. A 45-minute co-op run into a real world with terrain that withholds the easy answer, weather, a hazard, a manifest, and a score for what you took and what you left.
That’s the only date on this page that’s a commitment. Everything past it — M2b onward — is fully designed and not cancelled, but it is planned, not promised. The next commitment gets made at the M2a exit gate, on evidence from actually running it, not on a plan written in advance.
The honest caveat, stated once rather than buried: the finished game has six properties that make it recognisably this game rather than a generic co-op puzzle. G1 delivers four of them. The other two — the house is the factory (arrives with M3) and the factory’s customer is not you (arrives with M4’s Consignment) — are not there yet at week 30. So G1 is honestly the first real session, not the first real game. A good session, worth playing — just not yet the whole picture.
Where things actually stand today
This is what exists, right now, not a plan for what will exist.
- M0.1 — closed. Four crates, a headless server, two clients exchanging messages over a real wire protocol.
- M0.2 — closed, goal met in part. Five avatars replicate over
wss://against the deployed server; every join is credential-checked. The infrastructure half of the goal is proven; the “a friend actually plays” half is carried forward, honestly, because there was nothing to play yet. - M0.3 — in progress, not complete. Terrain, four block types (belt, splitter, lifter, sink) and the flow-graph analyser exist and work: cargo visibly moves from a vein to a planted Drop through blocks you place by hand. Still missing, and tracked as open work: the loadout-recovery rule, the three cargo classes, the run timer and score screen, the tuned wasm size measurement, and the five-friend playtest that actually decides whether this is fun. Read the devlog for how the pieces that do exist got built.
Nothing past this point is built yet. It’s designed — read the devlog and release notes as that changes.
The gates
| Gate | Milestone | A player can… | Status |
|---|---|---|---|
| G0 — The Probe | M0 | Place four block types on flat terrain, route cargo to a Drop by hand or by belt, race a clock with friends, see a score | In progress — the routing half exists; the clock and score don’t yet |
| G1 — The Expedition | M2a | Cross a gate into a real world with terrain that withholds, weather, and a score for what was left behind | The one committed gate — target week 30 |
| G2 — The Address | M3 + Consignment I | A house that’s a factory, running while you sleep, on a street you can hear, filling an order for someone who writes back | Planned, not promised |
| G3 — The Pipeline | M5 | Two worlds with different physics, and the first cross-world trade | Planned, not promised |
| G4 — The Place | M6′ | Sit still until fireflies land; cook with friends; log in with no goal | Planned, not promised |
| G5 — Public | M7 | A stranger arrives without any of us present | Planned, not promised |
G0 is a disposable probe by design — its whole job is answering one question: is building a rig together, under a clock, actually fun? If the answer is no, the honest thing to do is stop, not keep building.
Why the gate matters more than the calendar
Every big host in this industry hides its downtime, hides its real numbers, and calls a ticket queue “24/7 support.” We’d rather publish a roadmap that’s allowed to say “not committed” than one that quietly implies a date nobody internally believes. If that means this page looks less impressive than a marketing deck with six dates on it, that’s the trade, and we’re taking it on purpose.