Skip to main content
All CollectionsGetting Started
Glossary of Novelcrafter Terms
Glossary of Novelcrafter Terms

You may be hearing some words for the first time - here is a comprehensive list of terms related to Novelcrafter and writing with AI.

Updated over a week ago

You may be hearing some words for the first time - here is a comprehensive list of terms related to Novelcrafter and writing with AI. If there are any terms that you are unsure about, or cannot see on here, please let us know.

Act

Acts organize the high level structure of a story. Three Acts are a commonly used way of dividing a story into beginning, middle, and end. But there are other structures that use four, or five, or even twelve.

Action Menu

The Action Menu (three vertical dots) allows you to perform actions based on the object you're interacting with. For example, in a codex entry, the action menu allows you to change the color, clear a thumbnail, or delete the entry. Most of the objects in the Novelcrafter workspace have action menus, so keep an eye out for those magical dots!

Additions

Additions are a way of adding to a codex entry based on something that occurs in the story. Additions are entered via the slash "/" menu when you are in the Write interface.

You can find out more about additions here.

AI Context

The AI context refers to the all the information you send to the AI. Depending on the prompt, this can include anything from codex, snippets, scene summaries, and complete scenes. If you customize your own prompts, you can have even finer control over the AI context.

AI Edit (Expand/Rephrase/Shorten)

Also known as text replacement prompts, or laser tools, these tools allow you to use AI prompting to change prose previously generated. You can use the default/system prompts, or create your own.

AI Vendor

The provider of individual AI models - comparable to a mobile phone company. The cell phone is Novelcrafter, and EE/T-mobile/Vendor of choice is the AI Vendor. Most AI Vendors work on a pay-as-you-go model, where you top up your account (for as little as $5), and each time you "call" the AI (i.e. Send a message, use write, summarize a scene), you are charged based on the input and output tokens.

OpenRouter is one such aggregator of AI Vendors, allowing you to access many models in a single account.

You can find out more about AI Vendors here.

Beat, Scene Beat

A Beat is the description of a moment in a scene. A scene is usually made up of several beats. Using a series of beats in a scene allows you to control how your story unfolds.

You can find out more about beats here.

Chapter

Acts are divided into Chapters. The chapter is the main structural element that will be presented to your reader, who will remain blissfully unaware that there are overarching Acts to the story.

Chat

Chat is a common method to use when interacting with a Large Language Model. A chat consists of a series of user queries and LLM answers. A Chat knows about older questions and answers within the same Chat thread.

You can find out more about chat here.

Codex

A repository of the characters, locations, objects, lore, etc related to your story. You can store your entries for future reference, and/or use them to give context for the AI when sending requests via Chat, beats in Write, or custom prompts.

You can find more about the codex here.

Codex Category

(internally also called Codex Groups)

A way to group related codex entries together. This is for human-organization of the codex, and will not be provided to the AI in messages that you send.

Codex Type

The default grouping of a codex entry (if no categories are applied). If you are unsure where to put a codex entry, give it the type of "other".

Context

  1. Information given to an AI message, so that the output response is nearer to your desired output.

  2. Context window - The number of input tokens you can input into a message. This can range from 4k to 1,000,000, depending on the LLM.

Clone (a Model/Prompt)

A cloned model/prompt is identical to the original. If you have cloned a system prompt, then this cloned version is now editable.

You can learn how to clone a prompt here.

Fine Tuning

Fine Tuning involves presenting a model with training data (requests, answers) that results in a custom set of model weights that are introduced on top of a base model. For example, if you fine-tune a model to correct sentences to be in your author's voice, the fine-tune will use a subset of weights that are different to the base model. These custom weights were calculated during the fine-tuning process.

Genre

Genre is a label that tells readers what to expect from our stories. Genres include sci-fi, romance, detective, thriller. Each genre has many subgenres, for example: military science fiction, romantic comedy, cozy mysteries, and supernatural thrillers.

Global Entry

A global codex entry can be considered one that is always "turned on", i.e. It's context is always fed to the AI.

Examples of global codex entries would be the genre and style guide.

Labels (Markers, Tags)

These are human-facing tags to organise your scene. Learn more about them here.

LLM

Large Language Model, aka the AI.

A Large Language Model is a neural network that encodes information via language. Its architecture is analogous to the fine neural structures inside our brains, but it operates digitally on numbers rather than the electrochemical reactions of human brains.

Moderation (Internal/External)

Some makers of AI models don't want you writing anything that goes against their terms and conditions. To ensure you can't break these, the manufacturers add in layers of moderation.

External moderation is seen when using GPT models - you will get an error message. This is done before the message is sent to the LLM.

Internal moderation is built into the models by the makers, and may result in the model telling you that it is uncomfortable, or no message being returned. You will still be charged for these messages as they have been sent to the LLM.

See NSFW.

Nested Reference

These are a way to organise your codex, by linking entries. When you call the parent entry to the AI (done in Novelcrafter by mentioning the entry by its name or an alias), any nested references will also be brought up as context for your message.

Learn more here.

Novel

A Novel is the basic unit of story. A Novel can be part of a Series. A Novel is divided into Acts, each Act is divided into Chapters, and each Chapter can have multiple Scenes.

NSFW

NSFW stands for "not safe for work", and includes any content you wouldn't want your boss to see on your screen. Think of gratuitous violence, or anything sexual. The tolerance of a LLM to NSFW content varies, however as a rule, the GPT models can are strictest.

Outline, Create from

This feature allows users to paste in their story outline to be added as scene summaries.

Pantser/Pantsing/Discovery Writer

A writer who prefers to have minimal planning prior to writing. You may find yourself somewhere on this scale on a whole story level, or on a scene level (i.e. you don't plan how a scene will unfold, even if you know what should happen).

Authors who are pantsers include Stephen King, Dean Koontz, and George R.R. Martin.

Plan (i.e. Grid/Matrix/Outline)

The interface in Novelcrafter that stores the acts, chapters, scenes, and scene summaries of your story. There are three modes within the plan interface:

  • Grid - the default view,

  • Matrix - offers the most control of the parameters you view your story with,

  • Outline - a simplified view allowing you to concentrate on the scene summaries.

Learn more here.

Plotter/Planner

A type of writer who prefers to plan out their projects/novels before writing. There is no requirement of how much planning you have to do to fall into this category.

Authors who are plotters include JK Rowling, John Grisham, and Brandon Sanderson.

Point of View (POV)

The perspective through which the story is told. Common POVs are third-person limited, third-person omniscient, and first-person. Some Genres tend toward certain POVs.

Progressions

Prompt

The instructions you send to the LLM. These can be simple or complex. There are system prompts built into Novelcrafter to fulfil the basic requirements, however you can clone and edit these.

You can learn more about prompts/prompting here.

Scene

Chapters contain one or more scenes. A scene is usually unique to a character point-of-view or setting or time. For example, if you wish to have two different character POVs in a chapter, you can have one scene for each. You'll spend the first scene with one character and then change scenes to the new character POV. The same applies to changes in setting or time.

If a chapter has a single POV, location and time, it will generally only have one scene.

Scene Beat

See beat.

Scene Summary

A brief description of what occurs during the scene for quick reference. In Novelcrafter, scene summaries are used to give context to the AI model on what has happened so far in your story.

Section

A segmented area within a scene that you can put in notes, or hide prose from being seen by the AI. The perfect way to add a "note-to-self".

Learn more about sections here.

Series

A Series is a collection of Novels. In Novelcrafter, Series can share codex entries if the author chooses.

Snippet

A document stored in the sidebar that can be referenced when on any interface via the pin feature. Snippets can also be referenced in cloned, custom prompts.

Style Guide

A Style Guide is a codex entry where you can explain your writing style to the AI.

System Message

The system message is the highest level instruction you can present to the AI. Novelcrafter's prompts include system messages that focus the AI toward different tasks such as: chat, scene beat completion, and editing. You can customize these system messages in your own prompt.

System Prompt

The system prompts are the default in Novelcrafter. They cannot be edited, however you can clone the prompt in order to add/remove models.

The system prompt is designed to always work. There are system prompts for each type of prompt within Novelcrafter.

Temperature

A parameter of AI that determines how coherent or crazy/creative the model is. Usually, the nearer to 0 the temperature is, the less creative it is and for some models the closer it follows your instructions.

Token

Tokens are sequences of characters that LLMs convert into numeric representations, enabling the numeric processing of language at the core of large language models. Tokens can sometimes be entire words, but more often are pieces of words. For a general rule of thumb, it takes 100 tokens to encode 75 words.

Tokens, Input

The tokens used in your prompt.

Tokens, Output

The tokens that the AI generates as a result of your prompt.

Did this answer your question?