Lemon filed 2 patent applications in H1 2026, all in AI & Machine Learning.
The applications cover an AI-powered game creation system that enables users to build games through a natural language chat interface, combining large language models with diffusion models. This technology includes automated generation of visually consistent game assets from single prompts, allowing game development without traditional art skills.
Lemon's 2 AI & Machine Learning filings both center on making game creation accessible through generative AI, though each addresses a distinct part of the problem. The first application describes a system built around 2 cooperating agents, where one handles code and game parameters through a conversational interface while the other generates visual assets using diffusion models, with both operating together in a single pipeline that allows real-time iteration on either side. The second filing tackles a specific challenge within asset generation: rather than prompting each asset independently and risking a visually inconsistent result, the system derives a unified style description from a single user input and applies it across all characters, backgrounds, and environments before any images are produced. Together, the 2 patents cover both the architectural approach to combining language and image models in game development and the method for keeping the resulting visuals coherent with one another.
All data sourced from USPTO patent filings. Google Patents may take several weeks to index recent publications. If a link is unavailable, search for the patent number at USPTO Patent Public Search.