The Berkeley Statistician Building the First Instant Game Engine.
Lisha Li spent her academic career mastering the math of data. Now, she is using it to automate creativity. Rosebud AI has evolved from a synthetic face generator into the most advanced text-to-game platform on earth. With a simple chat prompt, thousands of creators are bypassing years of coding to build fully playable, high-fidelity browser games. In the world of Rosebud, the only limit on the virtual world is the user's vocabulary.

Lisha Li
Founder & CEO · Rosebud AI
The Democratization of Play
For decades, the barrier to entry for game development was a brick wall of C++ and complex physics engines. Lisha Li, a PhD from UC Berkeley, saw an opportunity to tear that wall down. Rosebud AI's mission is simple: if you can describe it, you can play it.
Rosebud's Gamemaker isn't just a tool; it's a creative shift. By using LLMs as a "translator" between human thought and game code, Rosebud allows anyone—from a 10-year-old student to a veteran designer—to prototype and deploy games in minutes. "We are moving from a world of consumers to a world of creators," Li notes. This vision has positioned Rosebud as a frontrunner in the emerging 'Creative AI' sector.
The Generative Game Stack
What sets Rosebud AI apart from traditional "no-code" builders is its deep integration of generative models. It doesn't just provide templates; it builds assets from scratch. Whether it's a pixel-art character, a synth-wave background track, or the complex Javascript logic for a platformer, Rosebud's AI handles the heavy lifting.
The platform utilizes the Phaser engine under the hood, ensuring that the games are high-performance and easily shareable in any browser. This technical choice allows for instant iteration—users can tweak game physics or change the entire art style with a single chat prompt, a process that used to take days of manual labor in traditional studios.
The Future of Social Gaming
As Rosebud moves into 2026, the focus has shifted from solo creation to social distribution. By making game creation as easy as tweeting, Rosebud is creating a new kind of "instant gaming" ecosystem. The platform now supports AI-driven NPCs that can hold conversations with players, creating dynamic experiences that are never the same twice.
With backing from Khosla Ventures and the YC network, Rosebud AI is well-funded to tackle the challenge of 3D world generation. As VR and AR become more mainstream, Rosebud’s ability to generate space and interaction via text will be the foundational infrastructure for the next generation of social virtual experiences.
"Generative AI will do for games what the smartphone did for photography. It turns every player into a creator and every idea into a world."
— Lisha Li, Founder & CEO, Rosebud AI
Company Timeline
- 2019
Lisha Li founds Rosebud AI and joins Y Combinator. Early focus is on AI-generated synthetic media and avatars for marketing.
- 2020
Launch of 'Generative Photos', an AI-powered library of diverse human faces, gaining massive traction in the creative industry.
- 2022
Strategic pivot toward gaming. The team recognizes that the same generative tech can be used to build entire virtual worlds.
- 2023
Launch of Rosebud AI Gamemaker in Beta. Users begin creating thousands of games using natural language prompts.
- 2024
Introduction of 'Pixie', an AI NPC assistant. Expansion of the platform to support multiplayer and complex 3D asset generation.
- Early 2026
Rosebud AI announces integration with major social platforms, allowing one-click publishing of AI-generated games to billions of users.
Frequently Asked Questions
Do I need to know how to code for Rosebud AI?
No. Rosebud is designed for non-coders. You provide text prompts, and the AI generates the code for you. However, advanced users can still view and edit the generated code to fine-tune their games.
What kind of games can I make?
Currently, Rosebud is optimized for 2D and 2.5D browser games, including platformers, RPGs, puzzle games, and top-down shooters. Support for 3D environments is currently in advanced testing.
Can I monetize my Rosebud AI games?
Yes. Rosebud AI provides hosting and sharing tools, and many creators are exploring ad-supported and 'play-to-earn' models as the platform expands its monetization features.
How does Rosebud handle copyright?
Rosebud AI uses proprietary models trained on permissive datasets. Users generally own the creative output of their games, though terms of service provide the platform with the necessary rights to host and display the content.
