Platform: Star (buildwithstar.com) | Genre: Voxel Sandbox | Status: Indev Pre-Alpha
The Review
There's real atmosphere in DEEPSMOKE. The dark, foggy voxel aesthetic actually earns the name. It feels moody and distinct rather than generic blocky-world lighting, which is a nice bit of visual identity this early on. The pixel-font title screen and the little progress-bar "Fund" tracker on the menu give it a scrappy, personal charm too.
The concept is the real hook though: drill with your goblin pal Brub, haul ore back to a vacuum to bank it before the deep reclaims your pack. There's a risk/reward loop built right into the premise, push deeper but don't get greedy, that a lot of voxel sandboxes don't bother setting up. That's a smart design instinct before you even touch the controls.
Playing it in its current state, it reads exactly as advertised, an indev pre-alpha roughly halfway down its roadmap. A couple of things stood out as areas that'll likely tighten up with more time. The menu text, especially on the main menu, could use a bit more size and letter spacing, it's a little hard to read at the moment, though that's a placeholder web font for now with a custom one planned down the line. On the plus side, it's already got mobile and console support built in, tested on Android and Xbox, along with controller rebinding as of the latest update, which is a genuinely strong feature set to have this early.
Bottom line, strong bones, a genuinely clever core loop, and clearly more early build than finished game. Which is exactly where you'd expect a two-week-old, halfway-through-the-roadmap project to be. What's genuinely impressive is that even at two weeks old, DEEPSMOKE is already one of the most advanced games in terms of 3D environment and world design on Star to date. The bones are strong here, and this one's shaping up to be a hit once it's finished.
In His Own Words: A Chat with the Developer
What stage is the game currently at? Indev pre-alpha. About halfway down the current roadmap.
How much time have you invested so far? Around two weeks.
Have you used any other AI tools to assist, besides Star? No, just Star and my own roadmap. I actually dropped Meshy and Tripo along the way and started doing my own models instead.
What made you choose Star over other AI game platforms? I tried them all, and none of them compared. It was partly timing, the jam, everything lining up, but mostly it came down to restrictions. Most platforms put a ceiling on what kind of game you could actually build, and I still haven't found that ceiling with Star. Now that the Ultra model is out, I barely use anything else. I get to focus on orchestration and asset creation instead of fighting the tool. Most AI tools sit there and guess at what you mean; Star's training data bridges that linguistic gap, and it's getting tighter with every update.
I've also been trying to build things for the community, not just for myself. Started with the Gamepad API, then hacked the leaderboard system into an async multiplayer data bus, passing bytecode commands between two game instances. Showed that to Ryze [Star Founder], and I was already on his radar from GOON [Another awesome game we'll be covering soon]. He's been a genuinely good founder in terms of actually listening to the people using the platform. It's a small community right now, but it's going to grow steadily. There's more platform stuff coming too, a map editor, a 3D model editor and painter.
I'm a 15-year Doom modder turned solo indie dev, thanks to Star.
Where did the original idea for DEEPSMOKE come from, and how much did it change during development? It came from years of Minecraft Realms sessions with friends, trying to recreate that same feeling but with my own spin on it. Not much has changed since the original idea honestly, just more of it has been written.
What part of the game are you proudest of, or most excited for people to experience? Probably the animal petting. The crater mechanic is a close second.
Was there a feature or mechanic that was especially fun or challenging to build with Star? Before the last server architecture migration, multiplayer worked by having every player's client take turns tracking where NPCs were, so they'd render consistently across everyone's game. That earlier era of multiplayer on the platform was genuinely interesting to work through. It ended up paving the way for what multiplayer looks like on Star today.
Did Star ever interpret one of your ideas in an unexpected way that ended up improving or changing the game? Not really, this build has been full steam ahead, single track.
What advice would you give someone using Star to turn a rough idea into a game that feels distinctly their own? Get a cheap notebook, fill it with doodles, rants, ideas, dreams. Take that, brainstorm, build a clear roadmap, and implement it. And if you get stuck on inspiration, live life a bit and let your subconscious do the heavy lifting. It knows what you actually want.
DEEPSMOKE is available to play now on Star. Follow crwlspaceINTERACTIVE for updates as the roadmap progresses, or check out more of his work on itch.io.