Getting tangled up in the beautiful landscapes of Cloud Gardens

Cloud Gardens takes players to an abandoned world, giving them seeds to plant to bring some natural beauty to the urban decay. By placing objects and seeds, players can encourage some striking growth and create some impressive dioramas.

Game Developer sat down with Thomas Vandenberg and Elijah Cauley, developers of Cloud Gardens, to talk about what interested them in creating a game about plants reclaiming an urban wasteland, the challenges that appeared when designing how players would guide plant growth, and the thoughts that went into creating this lost world.

What interested you about greenery growing over abandoned wastelands? What made you want to make a game about it?

Vandenberg: Starting with an earlier prototype titled “Garbage Country”, I was experimenting with the themes of players building stuff out of scrap and inhabiting abandoned places. The plants then were a way to add elements outside the player’s control to this inhabited space. A player returning later to something they built would not find it in exactly the same state, because plants would have overgrown their creation.

These elements lead to an aesthetic that is also found — and celebrated — in urban exploration photography. As the plant simulation got transplanted from the earlier prototype to Cloud Gardens, this genre of photography became our main source of inspiration; the goal being to let players build dioramas that end up looking like those real-life abandoned places.

What thoughts went into creating the gameplay of Cloud Gardens? In creating how players would control and interact with the plants/environments?

Vandenberg: Things started off with just a plant simulation that looked and felt fun: you throw a seed, it grows into a plant. I was looking for a way to turn it into a gamey game with goals and progression. It was a bit of a struggle coming up with traditional gameplay. I knew that I wanted players to place both the plants and the objects, creating as much as possible of the diorama themselves. Looking for a way to tie that into a progression loop, we decided on the simplest way to connect these elements: “objects make plants grow”. It doesn’t make much sense in a real-world interpretation, but it’s the most direct way to tie together the basic ingredients of the game.

Cauley: When I was brought on towards the beginning, Thomas had designed a really compelling plant growing simulation, almost exactly the one you see in the game today. However, back then the entire plant grew once you threw a seed. As we began looking for ways to turn this system into a gameplay loop which we could design levels and puzzle around, we realized we wanted to tie the player’s actions into the plants growth (as opposed to having it be automatic). At the same time, there was also a system in place for placing items in the scene around the plants.

If I recall correctly, I think one day I suggested to Thomas that we make the items the players were placing responsible for making the plants grow. From there, we found it was fun to have the plants grow as you filled out the scene, and continued to tune and balance that main gameplay loop.

How did you decide how much control to give to the player over the plant growth? Why did you decide on what you did?

Cauley: Similar to my previous answer, in the beginning we had plants growing fully all on their own. While this was fun, and something we’d return to in the Sandbox Mode, we ultimately decided that in order to design levels and puzzles, we’d need to give the player more control over the plant growth. I think we also intuited that there was something very satisfying about starting with a seed of a plant and having the actions you take grow it out fully.

Vandenberg: However, limiting the amount of control players have over the “shape” of the plant is intentional. I wanted players to be able to relax and watch the plants grow instead of supplying them with tools for deeply influencing the plant’s growth. I feared that the latter would lead to a more frustrating process of trying to get the plant to grow in a particular way.

Technical considerations come into play here as well. The plant simulation is pretty much just a brute force approach of deciding where branches should go. There are some parameters under the hood to influence this behavior. Though not impossible, adding player-controlled factors (like sunlight, water, nutrients) would push this simulation to the limit.

What drew you to the art style for Cloud Gardens?

Cauley: Thomas can speak to this better, as he’d already defined the core visual look of Cloud Gardens when I came on board. However, I can say that the art style of Cloud Gardens was one of the things that drew me to want to work on the project in the first place.

Vandenberg: It started out as an 3D extension of pixel art, which I’m more familiar with. I think pixel art is a nice way of stylizing that allows you to focus on the essence of what you’re drawing, and not get trapped into trying to approach AAA quality graphics. There’s also just something inherently ‘cute’ about pixel art 3D. And I found that this style is a good fit stacking/piling/building objects in random ways - the low level of detail makes it really lenient to ‘smushing’ objects together.

The campaign mode balances between "the natural and the manufactured." What ideas went into giving the game a campaign mode? In giving players a sort of structured challenge in growing plants?

Vandenberg: Once it was established that the goal of the game was to fully overgrow a scene, and that you’d do this by placing objects to make plants grow, we started thinking about how to string together these scenes in a type of campaign. Our initial ideas for the campaign were much more elaborate than the form that eventually ended up in the game.

We toyed with the concept of having the game world be one continuous world, that the player overgrows part by part, progressing each time a section is fully overgrown. However, I found that there were two issues with this initial approach. First, I did not think the aesthetics lend themselves very well to big worlds. I personally like the blocky nature of the graphics to be visible, and this is no longer the case at long distances. Second, it would limit the stuff we could do in individual “levels” because they would have to be tied to the larger world in a somewhat sensible way. In the end, we took a queue from games like LEGO: Builder’s Journey and ISLANDERS and opted for a more rinse/repeat type system where the levels are disconnected. I think this also makes for a more satisfying reward loop and session length.

Cauley: I think Thomas sums it up pretty nicely. I’ll also add that we wanted to give folks who wanted to engage with the plant-growing simulation, but felt lost or aimless in the sandbox mode, a way to engage with the game. Another little tidbit in relation to balancing the natural and manufactured, we had a rule when designing campaign levels that there couldn’t be any plants or ‘natural material’ (dirt aside) in the level before the player began planting seeds - that everything had to come from the player. We felt like it made it even more satisfying once they ultimately finished the level to know that everything in the scene they had planted.

What appealed to you about giving players a sandbox to freely grow their plants in? What drew you to add this mode as well?

Vandenberg: The sandbox was the original motivation of the game: plant some seeds and see what happens. I think it was clear from the beginning that this should be included. Sometimes I wondered how the game would’ve looked if it was only a sandbox and we had focused all of our efforts there.

Cauley: I agree with everything Thomas had to say. The sandbox always felt very natural in Cloud Gardens, and it mirrored the original fun of the prototype of throwing around seeds and watching them grow. Additionally, after we launched the game in Early Access, we loved seeing all the gorgeous gardens folks were making in our Discord. Their constant creativity and pushing the bounds of the sandbox mode definitely had a big impact in our continued support of the mode.

What ideas went into the creation of the locations you could grow plants over? Into the plants that you brought into the game?

Vandenberg: We did a lot of brainstorming about locations, and urban exploration photography is an endless source of inspiration. It often started with one photo that made us think “would it not be cool to create a scene that looks like this”. We ended up grouping these inspirations into themes (e.g. railway stations, rooftops) and picked the ones that appealed to us most. We had a bunch more on the whiteboard, but you can only do so much. Something similar goes for the plants, but there a lot more gameplay considerations that came into play. Things like: okay we need a plant that can cover a lot of horizontal ground easily, what kind of species could that be?

What ideas went into creating how the plants interacted with the environments?

Vandenberg: A big part of this is technological constraints. The plant simulation can handle a few types of interaction with the environment. Most importantly: the plants can either avoid or stick to objects. Plants that stick to (climb on) objects are just extremely satisfying to watch, so a lot of the game is based on that. More complicated ways of interaction (like nutrients or sunlight) would create the need to teach the player to use and exploit those mechanics. When they do, does that lead to the most beautiful dioramas? The fairly minimal interaction currently in Cloud Gardens gives players a lot of freedom to creatively solve the levels in a way that aesthetically appeals to them.

What drew you to create a moving diorama video of completed gardens? What do you feel that added to the game?

Vandenberg: I think it’s good for a game to provide ways for the player to ‘break out’ of the game - to share and admire what they did not just from inside the game. Those turntable videos are a crystallised form of your creation that people who don’t play the game can see as well. Much better than just a saved game. I would have liked to provide more sharing and interoperability here, because they can also be a really good marketing tool.

You mention that you have a Discord server with "a lively community of gardeners". What challenges did you face in building up a community around this game? What ideas and thoughts went into growing your audience for this game, in advance and now that it's released?

Vandenberg: I honestly think we were pretty naive in our approach here. We just set up a Discord with an invite link inside the game. I think it’s almost an expectation at this point that your game has at least a Discord community. There’s quite a few active players and truly skilled builders that hang out there. They’ve been a huge help during the game’s Early Access, and it has been a lot of fun to engage with the community in general.