Lovelace Studio is using generative AI to build Nyric, a tool that can help bring gaming worlds to life for player builders who work in harmony.
It's a toolkit for building a community-driven multiverse, or metaverse, as you wish. Players can use AI tools to craft their own worlds, bringing realms to life in a survival crafting sandbox. The aim is to give agency to solo creators and social gamers and then connect them together, said Kayla Commali, CEO of Lovelace Studio, in an interview with GamesBeat.
"We're doing generative worlds with AI and gaming," Commali. "We've been doing it for four years now since conceiving the idea. But lately it's become something pretty tangible, pretty exciting. It's a survival craft game for generative worlds, like Midjourney."
With Nyric, players can built generative AI worlds like Alice in Wonderland or Viking worlds, with any kind of style they want. It adds characters, themes and styles. Each realm is a hexagonal grid, interconnected with neighbors. In the gameplay loop, you can do pioneering and engage in regional diplomacy.
Companions agents, the faebots, are pioneering toolkits for player abilities in these hexgrid realms. Players are incentivized to create connections, such as forming trade routes with neighbors, leading community factions, and expanding their empires. The game worlds are persistent and built as procedurally generated hexagonal grids. Inside those worlds the players can craft the details and enjoy emergent gameplay.
A resource economy drives community-centric gameplay on one side, while influence and diplomacy powers individual competition on the other. Network effects take hold in a persistent 3D multiverse, where early discoverability can catapult creator virality.
There are a lot of AI game companies out there. But Commali said that Lovelace's approach to enabling players to create their own virtual worlds is different. Players can use prompts to generate interactive 3D survival craft realms with multiple biomes.
And inside the world is an AI companion that Commali calls a faebot. This faebot has personality traits and can remember things that have happened in the world. Players can modify the parameters of the world, including the sky, the ground, the environment or the weather. Soon enough, it looks like a unique place.
The virtual worlds also have multiple users and they can be linked. Players can discover each other, and form a social network. Players can craft things together in the world like players do in the world of Valheim.
After generating the world in a procedural way, players have to customize the world to make it unique. This means that it enhances human creativity, rather than doing all the work with AI and having the player take the credit for it.
You can customize your assets, the stories, and the themes of the world. Lovelace is testing the game on Steam shortly.
Commali and her Boston-based team have a background in robotics and computer vision. The team has raised $1.2 million from investors including Sequoia Capital Scout, HalfCourt Ventures, Blindspot Ventures, and Umami Capital. And it plans to raise more money.
The company has been at it for four years, and it is contemplating a move to Web3.
"There's a lot of external pressure to go to Web3. We haven't done it yet. Only if it makes sense. We've been really intentional about this," Commali said.
The team is planning to bring on an AI engineer and a tech artist to support development.
Over time, the company expects to integrate with the Unity game engine. And it helps players easily share their worlds in an open platform.
"We definitely didn't want it to be like a walled garden. The word sounds like an isolated experience. Our structure is like a representation of social networking. You scale and grow and there's no real ceiling," she said.
Origins
Commali has been on "self-taught journey" from biology to robotics and gaming. She learned computer vision. And cofounder Alex Engel has been in games for a couple of decades at places like Disruptor Beam and Turbine. Their team has grown to five people.
"We were working with a lot of telemetry, a lot of runtime data and simulation systems, and also working with this extensions, starting five or six years ago," Commali said.
The focus was on runtime compute at the consumer level that powers gameplay, and allows emergent gameplay to percolate. Then generative AI came along and the team pivoted into procedural generation.
"We believe a lot of people want to be content creators," Commali said. "About 65% of social media platform users say they want to be a content creator. Only 3% are."
So the idea became to provide the tools to make it easy for players to become more like developers. Rather than build its own large language models, Lovelace Studio is building its platform as an API on top of OpenAI.
"We have a pipeline that breaks down the world into like the biome, into the styles, the time periods, character and behaviors," Commali said.
