Settings

⚙️ Settings

Settings controls how the app looks, what the AI writes each turn, which AI it connects to, and some stored data. Your settings stay on this device.

To open it, select Menu on the main menu, then Settings. During a game, open the Menu and select Settings.

The Prompts and Tools tabs have their own pages: Prompts and Tools. This page covers the other four tabs.

How to Change the Narration Layout

1. Open Settings.
2. Open the Display tab.
3. In the Narration section, select Pages or Chat under Narration Layout.

The game changes at once. See Narration for what each layout shows.

How to Color Quoted Speech

1. Open Settings.
2. Open the Display tab.
3. In the Accessibility section, select the Quote Color checkbox.
4. To use your own color, set Light Mode Color or Dark Mode Color. The row names the mode on screen. Reset to Theme brings back the theme's color.

To set quoted speech in italic, select Quote Italic. It works with or without Quote Color.

How to Change the Narration Font

1. Open Settings.
2. Open the Display tab.
3. In the Accessibility section, pick a typeface in the Narration Font list.
4. To tune it, select Customize… beside the list.

The font changes the story text only. Use Global uses the app's Font.

How to Turn On a Thinking Mode

1. Open Settings.
2. Open the Output tab.
3. In the Reasoning section, select Inline, Planning or Staged under Thinking.

The line under the control says what the picked mode does. Native adds no thinking step. See Reasoning for the cost of each mode.

How to Limit Active Characters

1. Open Settings.
2. In the mode select next to the title, select Advanced.
3. Open the Output tab.
4. Under Thinking, select Staged. The Limit Active Characters row shows only in this mode.
5. Select the Limit Active Characters checkbox.
6. Type the largest number of entities to stage each turn. The default is 5.

How to Restore Default Worlds

1. Open Settings.
2. In the mode select next to the title, select Advanced.
3. Open the Data tab.
4. In the Storage section, select Restore Default Worlds. The button is off when you have deleted none of the bundled worlds.
5. Select Confirm in Restore Default Worlds.

Each deleted bundled world comes back at its latest version. Worlds you still have stay as they are.

---

Simple and Advanced

A mode select next to the Settings title shows Simple or Advanced. Open it to see what each mode does.

Mode	What it shows
Simple	The settings for everyday play: appearance, scene, narration, reading, the core turn passes, the endpoint connection and autosave
Advanced	Every setting, plus the Prompts and Tools tabs

The mode only changes what you see. A hidden setting still applies. A dot on the mode select tells you that a hidden setting is off its default.

In the tables below, Advanced marks a row that shows in Advanced mode only.

Display

What you see and hear.

Appearance

Setting	What it does
Theme	Light, Dark or System. System follows the light or dark setting of your operating system.
Theme Color	Sets the color palette for both themes. Preview theme… shows the palette on sample controls, and you can edit each color there.
Font	Sets the typeface for the whole app. Customize… tunes its Font Size, Bold Weight, Italic Slant, Line Height and Letter Spacing.

Scene

Setting	What it does
Background Music	Plays the music of each location during the scene
Location Background	Shows the location image behind the game
Background Fade	Fades the location image toward the background color, so the text is easier to read. Shows when Location Background is on.
Scene Images	Makes an image of each turn. See Scene Images. Your next action waits for the image, so each turn takes as long as your image server needs. Shows when Enable Image Generation is on.

Narration

Setting	What it does
Narration Layout	Pages shows one turn at a time, and the page buttons go back. Chat shows every turn in one list that you scroll, with your actions on the right.
Narration Reveal	Choose reveal animation… sets how each sentence appears as it streams: Fade, Move in, Scale and Blur, with an easing and a minimum speed. The dialog has a preview.
AI Language	Sets the language or style of the narration, the choices and the answers in Formaquestion. Pick a suggestion or type your own, such as formal English.
Paragraph Limit	Advanced. None, Single or Auto. Auto fits the paragraph count to Max Output Tokens, so the turn has a planned ending. With None, a long turn can stop at the token cap with no real ending.
Markdown Formatting	Advanced. Lets the narration use bold, lists and tables. See Text Formatting.

Accessibility

These settings change the story text only, not the rest of the app.

Setting	What it does
Narration Font	Sets a different typeface for the story text. The list has typefaces for dyslexia and low vision. Use Global uses the Font above.
Narration Text Size	Scales the story text
Line Spacing	Sets the gap between lines of story text
Quote Color	Colors text in double quotes, in the narration and in your own lines. Single quotes and apostrophes stay plain.
Light Mode Color, Dark Mode Color	Replaces the theme's quote color in the mode on screen. Reset to Theme removes your color. Shows when Quote Color is on.
Quote Italic	Sets quoted speech in italic, also on the choice buttons. It works with or without Quote Color.

Reset Size & Spacing sets Narration Text Size and Line Spacing back to their defaults.

Inspection

Advanced. These settings show work that the app usually does out of sight. They change nothing the AI writes.

Setting	What it does
Show Reasoning	Shows the model's private reasoning as a Thinking… note above the turn. The app saves the reasoning either way, so past turns show it when you turn this on.
Show Silent Requests	Shows background requests in the status bar and the context viewer: memory summaries, diaries and notes on new entities

Output

What the AI makes each turn, and what it carries forward. Most of these settings add or remove a request per turn. More requests give more features but make each turn slower.

Turn Extras

Setting	What it does
System Prompts	Three checkboxes for the requests that run after the narration: Choices, Stat Updates and Location Change. Turn one off and its feature stops, and its editor in the Prompts tab hides.
Move Automatically	Applies the move in your action before the scene is written, so the scene happens in the new location. It skips the Move to…? question. Shows when Location Change is on.

Reasoning

Thinking sets how the AI plans a turn before it writes it.

Mode	What it does	Cost
Native	Adds nothing. A reasoning model thinks as usual; other models answer at once.	One request per turn
Inline	Reasons privately, then narrates, in the same request	One request per turn
Planning	Plans the turn in its own request, then writes it. The most reliable mode for small models. Marked recommended.	One extra request per turn
Staged	A director picks who is in the scene. Each staged entity plans its motivation. A storyboarder makes the plan. The best mode for a consistent cast.	Several extra requests per turn. The slowest mode.

Turn on Show Reasoning to read what a mode wrote.

Setting	What it does
Limit Active Characters	Advanced, Staged only. Sets the largest number of entities the director stages each turn. Each staged entity adds its own request. With the checkbox off, the scene stages as many as it needs.
Native Reasoning	Advanced. Sets whether a reasoning model thinks, and how hard. The checkbox turns reasoning off. The list sets the effort, and Model Default sends no hint. Every prompt set to Global follows this row. A model with no native reasoning shows a note here in place of the control. A model that always reasons keeps the checkbox on, with a note.

Tools

Advanced. Tools lets the AI call Tools to get information it does not have. Each round of calls adds a request, so turns take longer. Only endpoints that support Tools get them. On other endpoints, the row shows a note in place of the checkbox. The Tools tab sets which Tools each prompt can use. See How to Turn On Tools.

Memory

Advanced. Memory Summaries, Semantic Memory, Memory Cap and Scene Recall decide what the AI keeps from older turns. Memory Settings lists each one with its default. To stop memories, see How to Turn Memory Off.

Semantic Memory downloads a small model of about 23 MB the first time you turn it on. A progress bar shows here while it downloads. Semantic Lore and Diary Recall use the same model.

Time

Advanced.

Setting	What it does
Time in Memory	Shows when Memory Summaries is on. Marks each memory with when it happened, such as day and time of day. Without it, the AI guesses how long ago things were.
Measured Clock	Lets the events of each turn set how much story time passes. Without it, each action takes the Time Per Turn. Adds one small request per turn.
Time Per Turn	Shows when Measured Clock is off. The story time each turn adds, as an amount and a unit (minutes, hours, days or weeks). The default is 1 hour. Enter 0 to hold the clock still.

See When Each Memory Happened.

Lore

Advanced. Semantic Lore activates dictionary entries by meaning, not only by keyword. Keyword activation does not change; this only adds entries. It uses the on-device model from Semantic Memory.

Characters

Advanced.

Setting	What it does
Describe New Characters	Writes a description for each entity the story invents. One extra request the first time a new entity is named. See How to Get Descriptions for New Entities.
Character Diaries	Staged only. Each entity present writes a diary entry as turns age out. Its recent entries shape its motivation. One extra request per entity.
Diary Recall	Adds older diary entries that fit the moment. It costs nothing extra. Shows when Character Diaries and Semantic Memory are on.

Choices

Continue the Story adds a [Continue the Story] button under the choices. Like any choice, it puts its text in the action box, and you send it. The story reads it as a push to keep going, not as something you do. Off removes it. On shows it with the choices. Always shows it also when the Choices request is off.

Attachments

Image Attachments adds an attach button to the action box, for up to 4 images with one action. The images go with that turn only. It also lets you attach images to a question in Formaquestion. Your model must read images; a text-only model returns an error. Each prompt's Include Attachments option sets which requests get the images.

Performance

Advanced. Concurrent Requests sends the choices, the stat updates and the location change at the same time, not one after another. Turns are faster on endpoints that handle parallel requests. Requests between turns, like memory summaries, also run together when it is on and one at a time when it is off. Turn it off if a local model with little memory slows down under the load.

Endpoints

Which AI the app connects to. The tab has its own tabs: Text, Image and, in Advanced mode, Tag Prompt. To set up a text endpoint, follow Connect Your Own AI.

Text

The Preset list holds your saved endpoints. Demo AI is the shared cloud endpoint, and you cannot edit it. On the desktop app, Built-In Engine runs a model on your PC. Add New Preset… makes a new one. Rename, Reset and Delete act on the preset you made.

Setting	What it does
Endpoint URL	The address of your model server. The line under it shows the full URL that requests go to. Trouble Connecting? opens a checklist.
API Token	The token for a hosted service. Leave it empty for a local server.
Model Name	The model the endpoint uses, exactly as the server names it
Context Window (tokens)	Advanced. How much the model keeps in context. Detect asks the server.
Max Output Tokens	Advanced. Caps how long each answer can be. It does not cap reasoning. Select Override Endpoint Limit to set it; without it, there is No Limit.
Sampling	Advanced. A section of five rows: Temperature, Repetition Penalty, Top-p, Top-k and Min-p. Each has a checkbox. An unchecked row sends nothing. Per-prompt values and built-in prompt values come before Temperature and Repetition Penalty.

Reset AI Endpoint, at the bottom of the tab, sets the URL, model name, token and limits back to their defaults.

Built-In Engine

On the desktop app, the Built-In Engine preset shows the engine's own panel in place of the fields. It has its own Simple and Advanced switch. Every row below Local Model is in the Engine section. See The Desktop Engine.

Setting	What it does
Local Model	Manage Models… opens the Local model dialog to download and load a model
Context Size	How much recent story the model can see. More context uses more GPU memory. Lower it first when a model does not fit.
GPU	Simple on the panel. Runs the model on the GPU. Off runs it on the CPU, which is slower.
GPU Layers, Layers	Advanced on the panel. Auto puts as many layers on the GPU as fit. Max puts all of them there and can run out of memory. Custom sets the count in Layers.
GPU Device	Which GPU loads the model. Auto uses your discrete GPU, or all of them when you have more than one, or the integrated GPU when you have none. All GPUs always splits a model across every GPU. Pick a GPU by name to use only that one. Shows while the GPU is in use.
Flash Attention	Advanced on the panel. Uses less GPU memory and is often faster. Turn it off only if an old GPU cannot run it.
Parallel Requests	Advanced on the panel. How many requests the model answers at the same time. More is faster, but each request gets less context and more GPU memory goes to them.
Temperature	How much randomness the model uses when it picks each word. About 0.7 fits most story models.
Max Output Tokens	Caps how long each answer can be. An answer at the cap ends at its last full sentence.
Top-p, Top-k, Min-p, Repetition Penalty	Advanced on the panel. Sampling limits on which words the model picks.

Sampling changes apply on the next turn. The other changes need Save & Reload Model. Reset to Defaults sets the panel back.

Image

Setting	What it does
Preset	Your saved image endpoints. You can edit each one, Default too.
Enable Image Generation	Shows the Generate with AI buttons. Off hides the rest of this tab and Scene Images.

The Connection section sets which server makes the images. The setup steps for each provider are in Image Generation.

Setting	What it does
Provider	ComfyUI, InvokeAI, Automatic1111 / Forge, NovelAI, or an OpenAI-compatible service on the desktop app. How to Set Up shows the steps for the picked provider.
Endpoint URL, API Token	The image server address and its token
Model	The checkpoint that makes the image

The Image section sets how each image is made.

Setting	What it does
Prompt Prefix	Quality and style tags put before every image prompt
Negative Prompt	Tags the image must not have
Portrait (W × H), Landscape (W × H)	Advanced. Image sizes. Portraits are for entities. Landscapes are for locations and thumbnails.
Steps / CFG	Sampling steps and how closely the image follows the prompt
Sampler	The sampling method
Face Fix	Automatic1111 / Forge and InvokeAI only. Draws faces again in a second pass. On Automatic1111 / Forge it needs the ADetailer extension on your server. On InvokeAI it about doubles the time.
Workflow (API Format)	Advanced, ComfyUI only. Replaces the default ComfyUI graph. How to Get This shows how to export one.
Board	Advanced, InvokeAI only. The InvokeAI board that gets the images
Qwen3 Encoder, Z-Image VAE, Anima VAE	Advanced, InvokeAI only. Shows for a Z-Image or Anima model. Leave them empty to pick one automatically.

Tag Prompt

Advanced. Shows when Enable Image Generation is on. It is the prompt that your text model gets to turn a description into image tags. Reset to Defaults brings back the original text.

Data

Saves, the Authoring Tour, and stored data. To back up your worlds, saves, library entities and library dictionaries, use Backup & Restore in the main menu's Menu. See How to Make a Backup.

Saves

Autosave saves after every turn, named by turn. It never changes your own saves. Autosaves Kept sets how many autosaves each world keeps; it starts at one, and a new turn replaces the oldest, and Re-generate replaces its turn's autosave. Lowering it shows a red warning while Settings is open: each world's oldest extras are deleted on its next autosave. Autosaves show under Load Game with an Auto tag. Autosave starts after the opening scene.

Authoring

Autosave saves your world while you edit it in the World Editor. The Save menu in the editor has the same checkbox. The setting applies to every world. See Saving and Discarding.

Autosave Pause sets how long the editor waits after your last edit before it saves. The range is 10 seconds to 5 minutes. The default is 30 seconds. A longer pause writes less often but leaves more unsaved. A large run of edits saves without waiting. The slider shows only while Autosave is on.

Start Authoring Tour opens the World Editor on a new world and shows an example for each step. See The Authoring Tour. Your other worlds do not change. This button shows only when you open Settings from the main menu.

Storage

Advanced. Each button asks you to confirm, and each is off when it has nothing to do.

Button	What it does
Restore Default Worlds	Brings back the bundled worlds you deleted. See How to Restore Default Worlds.
Clear Cached Images	Deletes the copies of linked images on this device. The app downloads them again when you are online. Your worlds and saves do not change.
Reset Tutorials	Shows the one-time tutorial notes again

Related

• 🔌 Connect Your Own AI: set up a text endpoint or the desktop engine
• 🧠 Story Memory: what the Memory and Time settings do to a story
• 🎭 Entities in Game: the entities the story invents, and their descriptions