NovelClaw  by iLearn-Lab

AI workspace for structured, memory-aware long-form story generation

Created 2 months ago
301 stars

Top 88.3% on SourcePulse

GitHubView on GitHub
Project Summary

A dynamic-memory-first AI framework, NovelClaw addresses the complexities of long-form fiction generation, chapter planning, and coherent narrative writing. It targets authors and builders who require robust continuity, iterative control, and inspectable workflows, offering a dedicated writing workspace that moves beyond single-prompt generation. The primary benefit is enhanced control and a clearer iteration path for complex, multi-chapter projects.

How It Works

NovelClaw treats long-form fiction work as an inspectable writing workspace, not a one-shot prompt wrapper. Its core approach organizes work into ongoing sessions, storyboards, manuscript surfaces, character/world views, and editable memory banks. This memory-aware system exposes execution state, logs, and chapter outputs, enabling direct control over chapter-level progress and facilitating revision and reuse across projects.

Quick Start & Requirements

  • Primary Install: Docker deployment is recommended. Use .\docker-start.bat (Windows) or ./docker-start.sh (Linux/Mac). Manual Docker setup involves copying/editing .env files with API keys and running docker-compose up -d. A one-click local stack script (.\START_LOCAL.bat) is available for Windows.
  • Prerequisites: API keys for LLM providers (configured in .env files) and Docker.
  • Access URLs: Local access typically starts at http://localhost:8010/select-mode (Portal), leading to http://localhost:8012/dashboard (NovelClaw workspace). A live platform is at colong-idea-studio.cloud.
  • Links: Live Platform, Local Preview, Docker Deployment Guide.

Highlighted Details

  • Memory-aware drafting with editable memory banks for continuity, revision, and reuse.
  • Inspectable runs featuring worker logs, progress traces, chapter outputs, and job detail views.
  • Persistent manuscript surfaces (storyboards, character/world views) kept visible as working material.
  • GitHub-safe deployment references and packaging, separating secrets and runtime data.

Maintenance & Community

Project authors are from Harbin Institute of Technology, Shenzhen. Documentation is available via Linux.do. No explicit community channels (e.g., Discord, Slack) or sponsorship details are provided in the README.

Licensing & Compatibility

Released under the MIT License, generally permitting commercial use and integration into closed-source projects without significant restrictions.

Limitations & Caveats

NovelClaw is explicitly designed not to be a one-shot prompt wrapper, requiring user engagement for its structured workspace. Its focus on inspectable, controlled generation may be more complex than desired for users seeking simpler output. Runtime databases and local keys are intentionally excluded from public release, necessitating careful local state management.

Health Check
Last Commit

1 day ago

Responsiveness

Inactive

Pull Requests (30d)
0
Issues (30d)
12
Star History
56 stars in the last 30 days

Explore Similar Projects

Feedback? Help us improve.