Flow steps
Each step of a flow has a type, which picks one of the kinds below, and
data, which holds that kind's settings. The entries come from the step
catalog the app plays and edits flows with, so the names and fields are
exactly what it accepts.
The look of a step
Every step except video also takes these fields in data. They are
left out of the tables below.
| Field | Type | What it is |
|---|---|---|
eyebrow |
string |
A short line above the title, such as a place or a date. |
background |
media ref |
The picture or video behind the screen. |
backgroundFill |
string |
A CSS gradient or color behind the screen, used when there is no background. |
vignette |
number |
How much the edges darken, from 0 (none) to 100. |
anchor |
"lower-left" | "centered" |
Where the text sits. |
accent |
string |
The color for highlights, as a CSS color. |
Questions and answers
- A question step needs
fieldon the node: the answer is kept under that key. ai_instructiontells the AI what the answer means. In a creation flow it goes to the AI with the answer. It holds up to 2,500 characters.- A choice's answer is the picked option's
id, or a list of ids withmultiple. The AI is sent the option's label instead. - A stat step's answer lists each stat's id, label and score, and the AI gets it whole.
Display steps
cover
Full-screen title over a background image. Called Cover in the editor. Shows something: no field.
| Field | Type | Required | What it is |
|---|---|---|---|
title |
string |
yes | The big title. |
subtitle |
string |
A line under the title. |
{
"id": "title",
"type": "cover",
"advance": {"mode":"timer","seconds":5},
"transition": "fade",
"data": {
"eyebrow": "Coriolis Station, day 4,211",
"title": "COLD STATIC",
"backgroundFill": "linear-gradient(160deg, #14233a 0%, #070a10 72%)"
}
}
The title card. Give it art with background, and pair it with a timer
on the node's advance to open a prelude.
statement
A heading and body copy over a background. Called Statement in the editor. Shows something: no field.
| Field | Type | Required | What it is |
|---|---|---|---|
title |
string |
The heading. | |
body |
string |
A few lines under the heading. |
{
"id": "silence",
"type": "statement",
"data": {
"title": "Nine days of silence.",
"body": "Telemetry stopped without a distress call. The relief crew never reported back.",
"backgroundFill": "linear-gradient(160deg, #2c2c34 0%, #0b0b0f 76%)"
}
}
reading
Long-form prose in a scrollable reading column. Called Reading in the editor. Shows something: no field.
| Field | Type | Required | What it is |
|---|---|---|---|
title |
string |
The heading. | |
body |
string |
The text. Long text scrolls in a reading column. |
{
"id": "log",
"type": "reading",
"data": {
"title": "Last log entry",
"body": "Day 4,202. The hum in the lower decks has changed pitch again. Okafor says it is the coolant. I don't think it is the coolant."
}
}
Its text is centered unless anchor says otherwise. Use it for a letter,
a log or an opening crawl.
video
A full-screen video clip (up to 60s). Called Video in the editor. Shows something: no field. The editor adds it with advance set to {"mode":"media"}; written without advance, it waits for the player.
| Field | Type | Required | What it is |
|---|---|---|---|
background |
media ref |
The video clip, with kind set to "video". |
|
anchor |
"lower-left" | "centered" |
Where the clip is anchored when it is cropped. |
{
"id": "n1",
"type": "video",
"advance": {"mode":"media"},
"data": {"background":{"url":"<video URL>","kind":"video"}}
}
The clip goes in background. Of the look fields above, it takes only
anchor.
Question steps
choice
A question with selectable options. Branches per option. Called Choice in the editor. Asks a question: set field. Can branch: one edge per option, when only one option can be picked.
| Field | Type | Required | What it is |
|---|---|---|---|
prompt |
string |
yes | The question, shown as the screen's title. |
ai_instruction |
string, up to 2,500 characters |
What the answer means, for the AI that makes something from the answers. Character creation sends it with the answer. | |
options |
object[] |
yes | The answers to pick from. |
multiple |
boolean |
Let the viewer pick more than one. The answer is then a list of option ids, and the screen can't branch. | |
optional |
boolean |
Let the viewer skip the question. | |
maxSelections |
number, whole number, at least 1 |
With multiple, the most options the viewer can pick. |
|
layout |
"list" | "cards" | "grid" | "select" |
How the options are shown: a list, picture cards, a grid of tiles, or a dropdown. | |
compact |
boolean |
Smaller tiles, for long lists of options. |
Each item of options: one answer the viewer can pick.
| Field | Type | Required | What it is |
|---|---|---|---|
id |
string |
yes | Saved as the answer, and matched by a branch edge's sourceHandle. |
label |
string |
yes | The option's text. |
image |
media ref |
A picture for the option. | |
ai_instruction |
string, up to 2,500 characters |
What picking this option means, for the AI. | |
info |
string |
More about the option, behind an info button. | |
referenceId |
string |
The reference id of a file this option stands for. The viewer can open the file. | |
fileTypeSlug |
string |
The linked file's type. | |
slug |
string |
The linked file's slug. |
{
"id": "approach",
"type": "choice",
"field": "approach",
"data": {
"prompt": "How do you dock?",
"layout": "cards",
"options": [
{"id":"quiet","label":"Run dark and listen first"},
{"id":"loud","label":"Announce yourself on all bands"}
]
}
}
To offer the world's files, give each option the file's reference id as
referenceId, with fileTypeSlug and slug, and use the reference id as
the option's id too. The player can open the file from the option; the
AI still sees only the label and the AI instruction.
shortText
A one-line free-text answer. Called Short Text Answer in the editor. Asks a question: set field.
| Field | Type | Required | What it is |
|---|---|---|---|
prompt |
string |
yes | The question, shown as the screen's title. |
ai_instruction |
string, up to 2,500 characters |
What the answer means, for the AI that makes something from the answers. Character creation sends it with the answer. | |
optional |
boolean |
Let the viewer skip the question. | |
placeholder |
string |
Faint example text in the empty answer box. |
{
"id": "salvage-prize",
"type": "shortText",
"field": "prize",
"data": {
"prompt": "What is the best thing you ever pulled from a wreck?",
"ai_instruction": "Give the character this item.",
"optional": true
}
}
longText
A paragraph free-text answer. Called Long Text Answer in the editor. Asks a question: set field.
| Field | Type | Required | What it is |
|---|---|---|---|
prompt |
string |
yes | The question, shown as the screen's title. |
ai_instruction |
string, up to 2,500 characters |
What the answer means, for the AI that makes something from the answers. Character creation sends it with the answer. | |
optional |
boolean |
Let the viewer skip the question. | |
placeholder |
string |
Faint example text in the empty answer box. |
{
"id": "navy-reason",
"type": "longText",
"field": "discharge",
"data": {
"prompt": "Why did you leave the navy?",
"ai_instruction": "Shapes the character's flaw and who is after them."
}
}
number
A numeric answer. Called Number in the editor. Asks a question: set field.
| Field | Type | Required | What it is |
|---|---|---|---|
prompt |
string |
yes | The question, shown as the screen's title. |
ai_instruction |
string, up to 2,500 characters |
What the answer means, for the AI that makes something from the answers. Character creation sends it with the answer. | |
optional |
boolean |
Let the viewer skip the question. | |
min |
number |
The lowest number allowed. | |
max |
number |
The highest number allowed. |
{
"id": "n1",
"type": "number",
"field": "age",
"data": {
"prompt": "How many winters have you seen?",
"ai_instruction": "The character's age.",
"min": 16,
"max": 90
}
}
rating
A 1–N rating scale. Called Rating in the editor. Asks a question: set field.
| Field | Type | Required | What it is |
|---|---|---|---|
prompt |
string |
yes | The question, shown as the screen's title. |
ai_instruction |
string, up to 2,500 characters |
What the answer means, for the AI that makes something from the answers. Character creation sends it with the answer. | |
optional |
boolean |
Let the viewer skip the question. | |
max |
number |
The top of the scale, which runs from 1. 5 when left out. |
{
"id": "n1",
"type": "rating",
"field": "confidence",
"data": {
"prompt": "How much do you trust the Guild?",
"ai_instruction": "1 is open hatred, 5 is blind loyalty.",
"max": 5
}
}
Stat steps
statFixedArray
Players assign a fixed set of stat scores. Called Fixed Array Stats in the editor. Asks a question: set field.
| Field | Type | Required | What it is |
|---|---|---|---|
prompt |
string |
yes | The question, shown as the screen's title. |
ai_instruction |
string, up to 2,500 characters |
What the answer means, for the AI that makes something from the answers. Character creation sends it with the answer. | |
optional |
boolean |
Let the viewer skip the question. | |
stats |
object[] |
yes | The stats to set, in order. |
values |
number[] |
yes | The scores the viewer hands out, one to each stat. |
Each item of stats: one stat the viewer sets.
| Field | Type | Required | What it is |
|---|---|---|---|
id |
string |
yes | The stat's key in the answer. |
label |
string |
yes | The stat's name, as shown. |
{
"id": "n1",
"type": "statFixedArray",
"field": "stats",
"data": {
"prompt": "Assign your ability scores.",
"ai_instruction": "Use these assigned stat scores exactly when filling numeric ability fields.",
"stats": [
{"id":"strength","label":"Strength"},
{"id":"dexterity","label":"Dexterity"},
{"id":"constitution","label":"Constitution"},
{"id":"intelligence","label":"Intelligence"},
{"id":"wisdom","label":"Wisdom"},
{"id":"charisma","label":"Charisma"}
],
"values": [15,14,13,12,10,8]
}
}
statPointBuy
Players allocate a configurable point pool across stat scores. Called Point Allocation Stats in the editor. Asks a question: set field.
| Field | Type | Required | What it is |
|---|---|---|---|
prompt |
string |
yes | The question, shown as the screen's title. |
ai_instruction |
string, up to 2,500 characters |
What the answer means, for the AI that makes something from the answers. Character creation sends it with the answer. | |
optional |
boolean |
Let the viewer skip the question. | |
stats |
object[] |
yes | The stats to set, in order. |
budget |
number |
yes | The points the viewer can spend. |
minScore |
number |
yes | The lowest score a stat can have. |
maxScore |
number |
yes | The highest score a stat can have. |
defaultScore |
number |
yes | The score every stat starts at. |
costs |
object[] |
yes | The total cost of each score, as a list of { score, cost }. A score not listed costs one point per step above minScore. |
Each item of stats: one stat the viewer sets.
| Field | Type | Required | What it is |
|---|---|---|---|
id |
string |
yes | The stat's key in the answer. |
label |
string |
yes | The stat's name, as shown. |
Each item of costs: what raising a stat to score costs in all.
| Field | Type | Required | What it is |
|---|---|---|---|
score |
number |
yes | A score a stat can reach. |
cost |
number |
yes | The points it costs in all to reach it. |
{
"id": "stats",
"type": "statPointBuy",
"field": "stats",
"data": {
"prompt": "Spend 27 points.",
"ai_instruction": "Use these scores exactly for the matching ability fields.",
"stats": [
{"id":"grit","label":"Grit"},
{"id":"wits","label":"Wits"},
{"id":"nerve","label":"Nerve"}
],
"budget": 27,
"minScore": 8,
"maxScore": 15,
"defaultScore": 8,
"costs": [
{"score":8,"cost":0},
{"score":9,"cost":1},
{"score":10,"cost":2},
{"score":11,"cost":3},
{"score":12,"cost":4},
{"score":13,"cost":5},
{"score":14,"cost":7},
{"score":15,"cost":9}
]
}
}
costs gives the total cost of reaching each score, not the cost of one
step: with the example, raising a stat from 8 to 14 costs 7 points.
statRollDice
Players roll dice totals and assign them to stats. Called Rolled Stats in the editor. Asks a question: set field.
| Field | Type | Required | What it is |
|---|---|---|---|
prompt |
string |
yes | The question, shown as the screen's title. |
ai_instruction |
string, up to 2,500 characters |
What the answer means, for the AI that makes something from the answers. Character creation sends it with the answer. | |
optional |
boolean |
Let the viewer skip the question. | |
stats |
object[] |
yes | The stats to set, in order. |
assignmentMode |
"assign" | "inOrder" | "perStat" |
yes | perStat rolls each stat on its own; assign rolls a set of totals the viewer hands out; inOrder rolls a set that goes to the stats in order. |
dieCount |
number |
yes | How many dice each roll uses. |
dieSides |
number |
yes | How many sides each die has (at least 2). |
keepCount |
number |
yes | How many of the dice count toward the total. |
keepMode |
"highest" | "lowest" |
yes | Whether the highest or the lowest dice are kept. |
rerollAtOrBelow |
number |
yes | Dice that land on this or lower are rolled again. 0 for none. |
modifier |
number |
yes | Added to every total. |
Each item of stats: one stat the viewer sets.
| Field | Type | Required | What it is |
|---|---|---|---|
id |
string |
yes | The stat's key in the answer. |
label |
string |
yes | The stat's name, as shown. |
{
"id": "n1",
"type": "statRollDice",
"field": "stats",
"data": {
"prompt": "Roll your ability scores.",
"ai_instruction": "Use these rolled stat scores exactly when filling numeric ability fields.",
"stats": [
{"id":"strength","label":"Strength"},
{"id":"dexterity","label":"Dexterity"},
{"id":"constitution","label":"Constitution"},
{"id":"intelligence","label":"Intelligence"},
{"id":"wisdom","label":"Wisdom"},
{"id":"charisma","label":"Charisma"}
],
"dieCount": 4,
"dieSides": 6,
"keepCount": 3,
"keepMode": "highest",
"assignmentMode": "perStat",
"rerollAtOrBelow": 0,
"modifier": 0
}
}
4d6, keep the highest 3, is dieCount: 4, dieSides: 6, keepCount: 3,
keepMode: "highest".