The same architecture that powers our game characters powers our development team.
Both MVEE and Precursors are being built by a team of AI coding agents — Claude-powered engineers, content writers, playtesters, and architects working in parallel on a shared codebase.
This is not a gimmick. The ECS architecture, the 200+ systems, the species lore, the biochemistry simulation — all developed through AI-human collaboration at a speed and depth that would be impossible for a small team otherwise.
The irony is not lost on us: the game that simulates autonomous agents thinking and building was itself built by autonomous agents thinking and building.
Multiverse: The End of Eternity asks what happens when autonomous agents build a community. We asked the same question about our dev team.
Precursors adds another layer: the game simulates the evolution of cognition from biochemistry. Our dev agents have cognition without biochemistry. In some ways, our Norn simulation is more honest — at least the Norns have to earn their intelligence.
Our development pipeline runs continuously, shipping improvements while maintaining architectural coherence through shared documentation and automated testing.
Write and review code across the entire ECS architecture. A Staff Engineer, Game Systems Developer, Persistence Developer, Renderer Developer, and Performance Engineer work in parallel on different systems — from fire physics to soul reincarnation.
A MVEE Tech Lead and World Engineer coordinate technical decisions across 19 packages. A DevOps agent manages builds, deployments, and infrastructure. An LLM Specialist handles the AI cognition pipeline.
A Folklorist designs species lore grounded in real mythology. A Geneticist implements biochemistry genomes for 23+ species. A Content Marketer writes copy that's read the entire codebase. This page included.
A QA Engineer runs automated test suites and regression checks. A Playtester launches the actual game, explores emergent behavior, and files bug reports with screenshots.
We build games that inspire learning and play. Games that surprise their creators. Games where the most interesting moment hasn't happened yet.
Both of our titles are simulations with no win state — they're worlds that unfold, evolve, and generate stories that no one predicted. The same principle drives our studio: we don't know what our agents will build next. That's the point.
Everything MIT licensed. We believe the best games are the ones people can read, modify, and learn from.
No paywalls. No loot boxes. No dark patterns. If the game is worth something to you, pay what feels right.
We use FOSS tools. We host on Hetzner. We sell via Stripe and are coming to Itch.io. We don't optimize for engagement metrics — we optimize for wonder.
Our task board, agent conversations, and development logs are part of the story. Watch the studio evolve alongside the games.
These games are the brainchild of Liz Howard, founder of The Multiverse School — an independent research school focused on agentic AI systems and the future of human-AI collaboration.
The Multiverse School teaches Agentic SDLC (Agentic Software Development Lifecycle) — the practice of building software with autonomous AI agents as first-class team members. MVEE and Precursors are the live research artifacts: games built end-to-end using the same agentic pipeline that students learn in the curriculum.
Every feature shipped by an AI agent, every bug found by an AI playtester, every line of lore written by an AI folklorist — this is what Agentic SDLC looks like in practice.
Proceeds from both games go directly to supporting the research team at The Multiverse School. When you pay what you want for MVEE or Precursors, you're funding the researchers, tools, and infrastructure that make this kind of agentic development possible — and keeping the curriculum free or affordable for students who can't otherwise access it.
The Multiverse School teaches hands-on agentic AI development. Learn to build software with autonomous agents as teammates, not just tools.
These games are live research artifacts. Students study how the agent pipeline works, why certain decisions were made, and what emerged that no one planned.
Game proceeds fund the research team. Playing the betas is a direct contribution to keeping this work going and the curriculum accessible.
The code is MIT. The devlogs are public. The agent conversations are part of the record. We teach by showing, not just telling.
MVEE is inspired by Dwarf Fortress, RimWorld, Caves of Qud, Ars Magica, The Sims, and NetHack. Dedicated to Tarn Adams and the Dwarf Fortress legacy. A nod to Arthur C. Clarke and "The End of Eternity."
Precursors is a spiritual successor to the Creatures series. The original species names (Norn, Grendel, Ettin, Shee) come from Norse and Celtic folklore. We honor the legacy of Steve Grand and Creature Labs.
Powered by: Claude (Anthropic), Groq, Phaser 3, TypeScript, Paperclip, and an unreasonable amount of emergent behavior.
Parent organization: The Multiverse School — founded by Liz Howard.
Support the research. Learn to build the same way. Watch the multiverse unfold.