Discover and explore top open-source AI tools and projects—updated daily.
RhythmicWaveAI-powered engine for long-form novel creation
Top 35.6% on SourcePulse
NovelForge is an AI-assisted long-form novel writing engine designed to tackle the challenges of maintaining consistency and control in large-scale creative projects. It targets authors and writers seeking a structured, yet flexible, environment for generating and managing complex narratives, offering a significant benefit in reducing AI-induced inconsistencies and streamlining the creative workflow.
How It Works
NovelForge employs a modular, card-based system where creative elements like characters, scenes, and outlines are defined and managed. Core to its approach are Pydantic-based "Dynamic Output Models" that enforce structured AI generation via JSON Schema, ensuring outputs adhere to predefined formats. A powerful "@DSL" allows for flexible, expression-based context injection into prompts, enabling complex data retrieval. For long-term consistency, a Neo4j knowledge graph tracks character relationships and plot dynamics, mitigating AI hallucinations. The system also features an "Inspiration Assistant" for conversational iteration and an "Ideas Workbench" for freeform brainstorming.
Quick Start & Requirements
pip, conda (recommended for environment management).npm.Highlighted Details
Maintenance & Community
NovelForge is described as being in an "early iteration stage," indicating active development. The project welcomes community contributions, suggestions, and bug reports via GitHub Issues. Specific community channels like Discord or Slack are not mentioned.
Licensing & Compatibility
The project is dual-licensed:
Limitations & Caveats
The workflow system is explicitly marked as experimental, with ongoing development needed for advanced features and robustness. The "ReAct" mode for the Inspiration Assistant is noted as a "rough" implementation potentially containing bugs. As an early-stage project, users should anticipate potential breaking changes and ongoing feature development. Enhancements to knowledge graph injection and workflow system capabilities are listed as future TODO items.
1 week ago
Inactive
merrymercy
Shengjia Zhao(Chief Scientist at Meta Superintelligence Lab),
google
grahamjenson
ThilinaRajapakse
google-research
triton-inference-server
tensorflow
visenger
PaddlePaddle