A systematic, context-aware framework for authoring coherent long-form novels and stories using AI assistance.
cordel is a structured writing methodology created by Alex Santos that overcomes the context decay problem in LLM-assisted storytelling. By dividing narrative production into modular world-bibles, character trackers, and outline stages, cordel guarantees flawless continuity.
From core premise to world-bible, beat sheet, chapter drafts, and continuity auditing.
Design choices, architectural patterns, and engineering decisions implemented for extreme reliability and developer experience.
Maintains structured markdown registries of character traits, relationships, knowledge states, and inventory, injected dynamically as targeted context.
Prevents meandering narrative by constraining each generation step to specific scene conflicts, character turns, and emotional beats.
Specialized auditor prompts cross-reference newly written prose against previous chapters to verify timeline integrity and character motivation.
Freely usable for commercial authors, game designers, and screenwriters seeking to integrate AI without sacrificing artistic cohesion.
Real-world usage, terminal configuration, and technical integration commands.
book-project/
├── 01_worldbible/
│ ├── characters.md # Knowledge states, arcs, voice guidelines
│ ├── setting.md # Geography, magic systems, technology rules
│ └── timeline.md # Canonical chronological events
├── 02_outline/
│ ├── master_beats.md # 3-act or 5-act narrative framework
│ └── chapter_01_beats.md # Granular scene goals
└── 03_drafts/
├── ch01_draft_v1.md
└── ch01_revised.mdKey clarifications regarding licensing, compatibility, deployment, and security.
Claude 3.5 Sonnet, GPT-4o, and DeepSeek-V3 excel due to their strong stylistic nuances and reasoning capabilities.
Yes. The world-bible and state tracking principles apply equally well to RPG game design and worldbuilding.
No special dependencies required; cordel works with standard Markdown files and any LLM interface.
By providing strict negative constraints and emphasizing physical sensory details over abstract exposition.