ForgeAI secures $95M to build a multimodal LLM tuned for game writing and dynamic narratives

AI · 6 min read

ForgeAI's multimodal model is trained on narrative structures, dialogue trees, and interactive storytelling datasets to produce context-aware NPC dialogue and branching plot content. The company claims the model excels at maintaining character voice over extended play sessions and managing continuity across emergent player actions.

The $95 million raise will fund model scaling, content moderation tools, and tools for authors to guide generative outputs with story blueprints and constrained style tokens. ForgeAI also plans to build plugins for popular game engines to simplify integration.

Investors included Benchmark and Epic Games Ventures, signaling strong industry interest in generative narrative tooling. ForgeAI will also invest in safety measures to avoid hallucinations and to ensure IP-safe content generation for licensed properties.

Game writers and narrative designers see a chance to prototype branching content quickly and to produce large volumes of variant dialogue with human oversight. ForgeAI emphasizes that its tools are meant to augment authors, enabling richer worlds without replacing creative control.