We present Dramamancer, an interactive narrative (IN) system that bridges authorial intent and audience input with natural language-based interactions powered by large language models (LLMs). Authors describe story settings in natural language and configure events via storylets, with natural language preconditions and contents. Players then experience these authored stories by inputting character actions in natural language, while the system adaptively generates story texts. LLMs serve as a narrative engine, dynamically incorporating both the authors' intent and the player's actions to flexibly generate stories at their intersections. By integrating LLMs with storylet-based IN authoring, Dramamancer aims to lower the authorial burden of creating INs while facilitating flexible player responses beyond predefined choices. We hope this demonstration inspires the UIST community to envision new types of interactive media that leverage intelligent technologies to bridge authors and audiences.
Wang et al. (Sat,) studied this question.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: