Constella: Supporting Storywriters' Interconnected Character Creation through LLM-based Multi-Agents

TL;DR

Constella tackles the challenge of interconnected character creation for long form storytelling by introducing a Large Language Model based Multi-Agent (LLM-MA) system with three features: FRIENDS DISCOVERY, JOURNALS, and COMMENTS. Grounded in formative interviews and deployed in a 7–8 day study with 11 writers, it demonstrates how multi-agent interactions can broaden writers' attention across character casts while preserving authorial agency. The work provides empirical insights into how AI driven character experimentation can serve as scaffolds for ideation, mindscapes, and relational dynamics, and discusses design implications such as deliberate constraints and intermediary materials to maintain writer control. Limitations include limited narrative progression in generated outputs and potential biases in prompts, pointing to future directions in enhancing generation quality, long form storytelling applicability, and cross cultural usability.

Abstract

Creating a cast of characters by attending to their relational dynamics is a critical aspect of most long-form storywriting. However, our formative study (N=14) reveals that writers struggle to envision new characters that could influence existing ones, balance similarities and differences among characters, and intricately flesh out their relationships. Based on these observations, we designed Constella, an LLM-based multi-agent tool that supports storywriters' interconnected character creation process. Constella suggests related characters (FRIENDS DISCOVERY feature), reveals the inner mindscapes of several characters simultaneously (JOURNALS feature), and manifests relationships through inter-character responses (COMMENTS feature). Our 7-8 day deployment study with storywriters (N=11) shows that Constella enabled the creation of expansive communities composed of related characters, facilitated the comparison of characters' thoughts and emotions, and deepened writers' understanding of character relationships. We conclude by discussing how multi-agent interactions can help distribute writers' attention and effort across the character cast.

Figures (15)

• Figure 1: Three main features of Constella that support storywriters’ interconnected character creation. (Left) FRIENDS DISCOVERY generates three characters in relation to a given character, facilitating character cast expansion. (Center) JOURNALS produces diary entries from multiple characters according to a shared theme, allowing the comparison of their inner mindscapes. (Right) COMMENTS enables characters to respond to each other’s journal entries, surfacing deeper layers of character relationships.
• Figure 2: Procedure of the formative study.
• Figure 3: Materials for the workshop. (A) Character List. (B) Character Cards. (C) Character Relationship Map.
• Figure 4: Constella’s default layout. The Left Sidebar opens the Journals Panel. The Right Sidebar shows a list of Character Cards, each of which opens the Profile Panel. The Create Button opens the New Character Panel.
• Figure 5: New Character Panel. (A) Add Attribute Button. (B) Save Button.