Summary

CharacterArc (弧光) is a desktop application designed for novelists, offering a structured, local-first workspace to manage project settings, character relationships, plot outlines, and chapter writing. It integrates AI collaboration tools, enhancing the creative workflow while maintaining full user control over data.

How It Works

Built on Electron, Vue 3, and TypeScript, CharacterArc utilizes SQLite for local data persistence, ensuring privacy and offline access. The architecture separates a main process (AI pipeline, SQLite management) from a Vue-based renderer process (UI). AI interactions are managed via a "Skill" system, enabling task-specific prompts and Agent Loop scheduling, supporting multiple AI vendors through OpenAI-compatible and Anthropic protocols.

Quick Start & Requirements

Requires Node.js 18+ and pnpm 10+. Installation involves pnpm install, with pnpm run dev for development and pnpm run dist for packaging. AI configuration necessitates API keys and endpoint URLs for supported text and image models. Primarily targets Windows, with minor adjustments for macOS/Linux.

Highlighted Details

• Local-First Data: All project data is stored locally in SQLite, guaranteeing user control and privacy.
• Project Isolation: Each novel project is managed independently, maintaining separate settings, chapters, and AI history.
• Chapter-Centric Workflow: Organizes content around chapters, featuring plot outlining, character/world-building management, and a TipTap rich text editor.
• AI Skill System: Employs a "Skill" system for AI task management, including Agent Loop scheduling and support for diverse AI providers.
• Visualizations: Includes a Cytoscape-based relationship graph for characters and organizations.
• Cover Generation: Offers a "Cover Workbench" for generating platform-specific cover prompts and preview images.

Maintenance & Community

The project acknowledges contributions and inspiration from "oh-story-claudecode" and "dama-cyber/Distilled-Novel-Toolbox". Community interaction is facilitated via a QQ group (https://qm.qq.com/q/lTQfy3AYvY).

Licensing & Compatibility

Released under the permissive MIT License, allowing for commercial use and integration with closed-source projects.

Limitations & Caveats

Primarily developed for Windows, requiring minor configuration changes for macOS/Linux. While data is local, AI functionality depends on external API access and user-provided credentials.