Intuition Check
A simple, private app for testing and training your intuition, on iPhone, iPad and Mac computers with Apple silicon. Practise your gut feeling with cards, numbers, or images — solo, or with a partner nearby.
Download on the App Store → Your own sets made with AI →Intuition
What intuition is
Intuition is knowing something without being able to say how you know it: a hunch, a gut feeling, a sudden certainty.
Psychologists describe one kind of it well — experience working faster than thought. Someone who has done a thing many times notices a pattern before they can put it into words: the feeling comes first, the words later.1
Many people also describe a second kind: knowing something that no experience could have told them — who is calling before the phone shows a name, or which card lies face down. Whether this kind exists is debated in science.23 This second kind is what the app lets you test, with values chosen at random.
How it can show up
Many people know moments like these:
- You think of someone, and a moment later they call.
- You pick up the phone to call a friend, and it rings in your hand — it is that friend.
- You wake up a couple of minutes before the alarm goes off.
- You know who is calling before you look at the screen.
- You remember someone you haven't seen for a long time, and they call that same day.
- Your first impression of a person turns out to be right.
- The answer to a problem comes by itself — in the morning, in the shower, on a walk.
- You feel someone looking at you, turn around, and someone is.
In telephone surveys of randomly chosen people in England and in California, between 65% and 78% said they had called someone who said they had just been about to call them, and 47–49% said they had known who was calling with nothing to go on.45 Most adults believe they can feel when someone they can't see is looking at them.6
Some of these have a known explanation. The body can keep track of the time you expect to wake: in one experiment, a hormone that prepares the body for waking began to rise about an hour before the time people had been told they would be woken.7 A pause can help with a problem: across many experiments, people solved problems better after setting them aside for a while,8 and in one study a night's sleep more than doubled the share of people who found a hidden rule that made a task faster to solve.9 And a first impression can be accurate: strangers who watched a teacher for 30 seconds in all, with the sound off, rated them much as the teacher's students did at the end of the term.10
Once, twice — or again and again
Once or twice, a coincidence like this is easy to put down to chance. When it happens again and again, it is natural to wonder whether there is more to it.
Memory is a poor counter here. It keeps the times the phone rang and forgets the many times you thought of someone and nothing happened. And with enough days, people and thoughts, striking coincidences are bound to happen now and then.11
That is why it helps to count. In this app, the chance of guessing right by luck is known exactly: it depends only on how many values can be drawn. The statistics show whether your results stay within what chance would give or depart from it — see Reading your results.
Intuition in everyday life
Feelings and signals from the body take part in many decisions. People who simply chose the poster they liked were more satisfied with it three weeks later than people who had first been asked to list their reasons for liking or disliking each one.12 In another experiment, people played a card game with four decks: two decks paid off over time and two lost, but the players weren't told which. The body reacted before each pick, and these signals leaned different ways in different people: on the whole, toward the winning decks for some and toward the losing decks for others. The researchers gauged how well each person sensed their body's signals by how accurately they could feel their own heartbeat. The more accurate that feeling, the more closely the person's choices followed the body's signals — for better or for worse.13
Intuition is most reliable where there are real patterns to lean on and quick, clear feedback. A strong feeling of certainty is not, on its own, a sign of being right.1
So being in good contact with your intuition means two things: noticing it and learning when it can be trusted. To learn when it can be trusted, note your state as you play: the statistics then compare your rounds by state and by time of day, so you can see when you do better. See Your state.
Is there more to it?
Science has not settled whether people can know what no experience could tell them. In 2018 a review in the main journal of the American Psychological Association concluded that the evidence for such effects is about as strong as for findings accepted in psychology.2 A reply in the same journal argued that such effects are impossible and that the results must have other causes.3 The question stays open, and the app lets you look at it with your own rounds.
How the app works
A round
Pick a guess set — cards, numbers or a set of pictures — and try to sense a value before it is shown. Then reveal it and see whether you were right. The app counts your hits and compares them with what pure chance would give.
Passive and active modes
The app has two modes:
- Passive intuition — the value is already chosen and hidden. You try to read what it is.
- Active intuition — nothing is chosen until you reveal the value. You try to will a value into being.
The statistics count the two modes separately, so you can see which one suits you better.
Three ways to play
The switchers at the top of the main view set the play style and the mode. The same choices are in Settings → Play style and mode.
One by one
One value at a time. Tap to reveal it, tap again to hide it, then reveal the next one. When Reveal without hiding is on in Settings → One by one, each tap shows the next value straight away. Show hint adds a line of guidance under the value.
With the guess tracking Method set to Enter your guess — as it is at first — the app asks for your guess before each reveal and scores the round itself. See What gets recorded.
Quick pick
All the values of the set are laid out in one grid. Each tap is a round of its own. In active mode the value is drawn at the moment you tap; in passive mode it is drawn in advance, right after your previous tap. A hit can be marked with a green flash and a buzz.
Number sets with more than 1,000 values have no grid: narrow the range in the set's settings, or play One by one.
Series
In active mode, choose the value you want, will it and tap it: the app draws a series of values and counts how many of them match the one you tapped. A series can be 2 to 12 draws long, or 100, 1,000, 10,000, 100,000, 1 million or 10 million. You set the length on the main view. In Settings → Series, the Series time slider spreads the draws over up to 10 seconds; at zero there are no pauses. Each series goes into the statistics as one entry.
Choosing Series switches the mode to active intuition.
Cards and numbers
A card set uses a deck of 36 or 52 cards, with or without jokers. A number set uses numbers of up to 9 digits — you choose the maximum number of digits. You change both in the guess set's settings view: on the main view, tap the gear beside the set switcher.
Clear view
Touch and hold anywhere on the main view to hide the switchers, the toolbar, the state bar and the hint. Touch and hold again to bring them back. The round counts and the result stay visible.
In Settings → Appearance and feedback, under Clear view, Show button adds an eye button to the toolbar — tap it to turn clear view on — and Remember clear view keeps clear view on between launches.
Round counts
The main view can show how many rounds you have played and how you are doing: one count and one result for each combination of mode and play style, with a button that resets the count and clears the result. Switching to another guess set also resets all the counts and results.
To keep the round counts and results between launches, turn on Save round counts and results in Settings → Statistics. It is available while Save statistics on device is on.
Guess sets
A guess set is what you guess from: cards, numbers or a set of pictures. At first the main view shows only Numbers and Cards. You can add the other sets that come with the app by tapping beside the set switcher on the main view.
The set switcher
The switcher at the top of the main view shows your guess sets. With the default Switcher style it shows buttons when it has one or two sets to show, and from three on the name of the current set, which opens a list of the sets. With Separate view chosen in Switcher style, the switcher always opens this list.
Each set has its own Use switch. A set with this switch off is offered nowhere — not in the set switcher on the main view, not in the partner game, not in the statistics — but it stays in your list of sets in Settings → Guess sets → Manage guess sets, and turning it back on brings its statistics back.
The button beside the set switcher opens Add a Set. It lists the extra sets that come with the app, ready to be added to the main view; below them are a button for creating an empty set, to which you can later add your own pictures, and the Your AI Assistant block for adding a set an AI assistant made. The gear button beside the set switcher opens the settings of the currently selected set. Both buttons can be hidden in Settings → Guess sets with Show add-set button and Show set settings button.
Ready-made sets
Tap beside the set switcher on the main view — or open Settings → Guess sets → Manage guess sets, tap and choose Ready-made sets. Open a set and tap Add to my guess sets.
The set you add is your own copy: you can rename it, reorder its pictures, leave some of them out of the game, play it with a partner or send it to someone.
A set from your own pictures
From the main view: tap beside the switcher, then Empty image set below the ready-made sets. Keep the name My images or type another and tap Create. The app switches to the new set and opens it.
Or open Settings → Guess sets → Manage guess sets, tap and choose Empty image set, then name the set and tap Create. The new set opens.
Inside the set, tap and add pictures from Photos or Files — several at a time — take them one by one with the Camera, or paste a copied picture or SVG with Paste from Clipboard.
Picture names
Tap a picture in the set's own view to see it larger. Its Name is the field under it: type one, change it or clear it. A picture added from Files starts with the file's name without its extension; one from Photos, the Camera or Paste from Clipboard starts with none. In a set with names in several languages, Names by language under the field holds the picture's name in each of them. Below the picture you see the format and size of the file the app keeps.
An AI assistant can make pictures and whole sets for you: Your own sets made with AI.
Order, layout and what is in play
In the set's own view, Arrange lets you drag the pictures into a new order and choose Images per row, Alignment and Scale to fit for the grid.
Touch and hold a picture for more actions:
- Don't use in game keeps the picture in the set, but it is never drawn. Use in game brings it back.
- Mark as “?” placeholder shows that picture while a value is still hidden in One by one. It is never drawn itself. A set can have only one “?” placeholder.
Renaming and deleting
The set's name is the first field in its view. In ready-made sets, editing it changes the name in the app's current language; the names in other languages are under Names by language.
To delete a set, swipe it in your list of sets in Settings → Guess sets → Manage guess sets, or touch and hold it, and choose Delete. The app asks first. The set's statistics are deleted with it.
Sending sets
Send a set
A set of your own pictures can travel as a single file with the extension .guesset. The file holds the pictures, their order, the layout and the set's names, as well as which pictures are in play and which is the “?” placeholder. Cards and Numbers are built into the app and can't be sent.
- Open Settings → Guess sets → Manage guess sets.
- Touch and hold the set and choose Export…. The share button inside the set does the same.
- Choose how to send it: Mail, Messages, AirDrop, Files or another app.
To play a set with someone nearby you don't need a file: see Playing with your own pictures.
Open a set someone sent you
Tap a .guesset file in Mail, Messages or Files, or accept one sent by AirDrop, and open it with Intuition Check. It becomes your own separate copy and appears at the end of your list and in the switcher.
A file already saved on the device can also be imported from inside the app: in Settings → Guess sets → Manage guess sets, tap and choose Import from file.
Partner game
How it works
Two devices play together: one person picks a value, the other tries to guess it. The devices talk only to each other (device to device when they are close, or over a shared Wi‑Fi network). Partner-game rounds don't go into the statistics.
The connection
Turn on Wi‑Fi on both devices. On the same Wi‑Fi network the connection works best, and the devices can then be anywhere on it. With no shared network they connect device to device — keep them close together.
The devices look for each other only while the partner game is open on both of them.
Start
- On both devices, open the partner game: tap the two-person button in the toolbar of the main view, or tap Open partner game in Settings → Partner game. If the button isn't there, turn on Show partner game button in Settings → Partner game.
- Each device shows Looking for a partner nearby…, then the other device under Seen nearby:. They connect by themselves.
The first time, your device asks whether the app may find devices on your local network. Allow it: otherwise the devices can't find each other.
A round
- Under Play with:, pick what to play with — any of your guess sets in use.
- Choose a role: I'll choose a card (or a number, or an image) or I'll try to guess. Your partner gets the other role.
- The chooser picks a value, the guesser makes a guess. Both devices then show the chosen and the guessed value, with Correct! or Not this time.
- Play again takes you both back to the first step, to pick what to play with and the roles for the next round.
Playing with your own pictures
When you play with one of your picture sets, your partner is asked to Download it first, unless they already have the same set. Their copy is used only in the partner game and never changes their own sets. Only the pictures they don't have yet are sent.
With Keep received image sets on in Settings → Partner game, the next game with the same partner starts without waiting. With it off, received sets are deleted when you leave the partner game. Received sets are listed in Settings → Guess sets → Manage guess sets under Received from partner, where Add to my guess sets makes a copy that is fully yours.
Your partner's name
On iPhone and iPad, apps see only a general device name such as “iPhone”, so two devices can look the same. Once you are connected, tap the pencil next to your partner's name and type a name of your choice. It is shown only on your device.
If the devices don't find each other
- Open the partner game on both devices, with Wi‑Fi on.
- Put both devices on the same Wi‑Fi network if you can; otherwise keep them close together.
- Guest and public networks — in hotels, cafés, offices — often block traffic between devices. Try another network, or no network at all with the devices side by side.
- If you declined the local network request, turn access back on in the Settings app: Privacy & Security → Local Network, then switch on Intuition Check.
- If a device appears under Seen nearby: but never connects, leave the partner game on both devices and open it again: the search starts afresh.
- Only two devices play together at a time. A third one keeps looking until one of the pair leaves.
Statistics
What gets recorded
Quick pick and Series always record your rounds. One by one records them according to the Method setting under Guess tracking in Settings → One by one:
- Enter your guess — the default. You pick your guess before the reveal, and the app scores the round.
- Mark yourself — after each reveal, tap ✓ or ✗. A reveal you skip without marking isn't counted.
- Off — One by one records nothing.
A guess set with only one possible value records nothing. Partner-game rounds are never recorded.
Looking at your statistics
Tap the statistics button in the toolbar of the main view, or tap View statistics in Settings → Statistics. The statistics open for the guess set, play style and mode you are playing. The filters at the top narrow them by Period, Guess set, Play style, Mode and — while you note your state — State.
Sections you don't need can be switched off in the statistics settings view — open it with the settings button in the statistics view, or from the main view through Settings → Statistics.
What the figures mean: Reading your results.
Keeping your statistics
At first, statistics are kept in memory only, for the current session: they are cleared when the app quits, including when iOS closes it in the background. To keep them, turn on Save statistics on device — at the top of the statistics view or in Settings → Statistics. The rounds of the current session are saved too. Rounds from earlier app sessions that were never saved can't be brought back.
Turning the switch off deletes the saved statistics. The app asks first, with Stop saving statistics?; after you tap Delete saved statistics, only the rounds played since the app was last opened remain, and only in memory: they are gone once the app is closed. This can't be undone.
Clearing
At the bottom of the statistics view:
- Clear all statistics for “…” — all statistics recorded with the guess set picked in the filter.
- Delete the filtered statistic records (count: …) — only the records the filters show.
- Clear all statistics — everything.
The app asks you to confirm each of these deletions, and none can be undone.
Export
Tap the share button in the toolbar of the statistics view, then Export statistics to a text file. The file is a table (CSV) of exactly the records your filters show: one row per round, or one per series in the Series play style. Spreadsheet apps such as Numbers can open it. You may also find it useful to show this file to your AI assistant and ask, for example: “In what circumstances does my intuition seem to work better or worse?”
Reading your results
The statistics don't just count hits. They compare your results with what random guessing would score with the same number of possible values, and say in plain words whether you are ahead of chance, behind it or in line with it.
Your result and chance
Under Correct guesses, Your result stands beside Random would get. The same 30% can mean opposite things: with 3 possible values, random guessing gets about 33%, so 30% is a little below chance; with a 52-card deck it gets about 2%, so 30% is far above it. The chance figure uses each round's own number of possible values.
The verdict
The app gives a verdict: from Consistent with chance through slightly and notably above or below chance, up to Far above chance or Far below chance. It also says how rare such a result would be for random guessing, for example “about 1 in 300 times”.
When rounds are few, the results swing widely by chance alone. So the verdict grows steadier with each new round.
Passive and active intuition compared
When you have rounds in both modes, Passive vs active intuition shows the hits and the hit rate of each, so you can see which one suits you better. It is not shown for Series.
Best streaks
Best streaks lists your longest runs of correct guesses in a row, with the day and time and the state you had noted. A pause longer than Max pause between rounds — you set it in that section — ends a streak, and so does a change of your noted state.
Over time
Hit rate over time draws your daily hit rate once you have rounds on two different days. By time of day shows the morning, afternoon, evening and night separately.
Values
When the statistics are shown for a single guess set with up to 100 values, they list the values you guessed right most often and the values you pick as your guess most and least often — including values you never pick, once you have made as many guesses as the set has values.
People often lean toward some values and avoid others, and some values may be guessed right more often than others.
Results can shift
Your results may change with your state and your circumstances. None of the following is a settled fact — they are possibilities worth noticing as you play:
- Your state. You might do differently just after waking or late at night, rested or tired, calm or stressed. Note your state as you play, and the statistics compare your rounds by state and by time of day.
- Confidence. Feeling sure of yourself, or full of doubt, could line up with how you score.
- Below chance counts too. Guessing right noticeably less often than chance would predict is also a departure from chance. One reading is that intuition is working against you at that moment, rather than that nothing is happening.
- Values. You might guess some values right far more often than others, and some you might never get right at all, even though you pick them just as often as the rest.
What the numbers can tell
The app can tell you whether your results depart from pure chance — never why, even when the reason seems obvious from the data. Treat the figures as something to explore, not as a fixed measure of yourself. See also Intuition.
Your state
The state bar
The bar on the main view lets you note how you are right now: Rested or Tired, Calm or Stressed, Confident or Doubtful, Focused or Scattered, your mood, and circumstances such as Just woke up, After coffee or After exercise. What you note is recorded with every round you play from then on, until you change it or it is forgotten.
The statistics then compare your rounds by state and by time of day, and each streak shows the state you had noted. See Results can shift.
When it is forgotten
Reset noted state in Settings → Your state decides:
- When the app starts — the default. The next launch starts with nothing noted.
- After a pause — forgotten once you haven't played for longer than the Pause you choose, from 15 minutes to 1 day.
- Never — kept until you change it.
Your own states
Settings → Your state → Manage states lets you rename, reorder, add and hide states, add axes of your own and switch an axis off with its own Use switch: a switched-off axis leaves the state bar, new rounds don't record it and the statistics leave it out.
To remove the bar altogether, turn off Note your state during rounds; then no state is recorded at all.
The bar is hidden when clear view is on, and in One by one when guess tracking is off.
Appearance and features
Themes
Settings → Appearance and feedback has four themes — Sand, Ocean, Forest and Dusk — each with a light and a dark version. Appearance offers Light, Dark and System; with System the app follows your device's setting and switches along with it. The theme also colours the playing cards and most of the ready-made pictures.
Feedback
Under Hit feedback, Vibrate and Animate mark a correct guess with a buzz and a green flash; either can be turned off. They work in Quick pick for every tap, in Series when every draw of a series hits, and in One by one when you enter your guess before the reveal. A miss gets neither the buzz nor the flash. Vibration also needs System Haptics on in the Settings app, under Sounds & Haptics.
Pulse animation on tap gives a brief pulse to the cell you tap in Quick pick and Series, and to the value you reveal in One by one when you mark rounds yourself or don't track them.
An uncluttered main view
Almost everything on the main view can be hidden. Each switch is in the Settings view it belongs to:
- Settings → Play style and mode: two Show switcher switches, one for the play style and one for the mode.
- Settings → Guess sets: Show switcher, Show add-set button, Show set settings button.
- Settings → Quick pick: Show grid options button.
- Settings → One by one: Show hint.
- Settings → Series: Show series length on main view, Show series time on main view, Show grid options button (the same switch as in Settings → Quick pick).
- Settings → Your state: Note your state during rounds.
- Settings → Statistics: Show statistics button.
- Settings → Partner game: Show partner game button.
- The settings of each play style (Quick pick, One by one, Series): Show rounds played, Show Results switch, Show result.
To take one guess set off the switcher, turn off Use in that set's settings; the set then leaves the partner game and the statistics too. Turning on clear view hides most of the main view's elements at once.
Features
Settings → Features has a switch for Series play style and one for Round counts and results on the main view. Switched off, a feature leaves the main view and Settings.
Language
Settings → Language lets you set the app's language independently of your device's language, and it sets how numbers and dates are written. Under Suggested Languages, the list shows the system language first, then each of your device's languages that the app offers, in the order set on your device; all other languages follow under Other Languages. When the new language reads in the other direction, close the app and open it again after switching (the writing direction switches correctly only after a restart).
Below the Language row in Settings is the Show hints in Settings switch; it shows or hides the explanations under the settings.
Free hour and subscription
The free hour
The app is free to download and includes one hour of in-app time to try everything in it. The hour is counted only while the app is open on the screen, and it adds up across all your sessions — no clock runs while the app is closed or in the background.
When the hour is used up, the subscription view opens by itself. If that wasn't enough time to decide about subscribing, you can take more: Reset the free hour gives you another hour, up to 3 times — 4 hours of free time in all.
The subscription
To keep using the app after that, there is a subscription with two plans, Monthly and Yearly. The subscription supports Family Sharing.
Subscribe
Open Settings → Subscription → Subscribe.
Manage or cancel
Subscriptions are billed by Apple. Settings → Subscription → Manage subscription takes you to yours. The app's subscription can also be managed the standard way, through the subscriptions of your Apple Account in your device's settings. Cancelling a subscription stops future automatic renewals, but you keep access until the current period ends.
A monthly subscriber can move to the yearly plan with Settings → Subscription → Switch to yearly plan.
Refunds
Refunds are handled by Apple: request one at reportaproblem.apple.com. We have no access to your payment details and can't issue a refund ourselves — but do write to us as well if something went wrong, because that is the only way we find out about it.
Restore a purchase
On a new device, open Settings → Subscription → Subscribe and tap Restore purchases. This also brings back a subscription that a family member shares with you. Use the Apple Account that subscribed, or one in the same family group.
Your data
In short
There are no accounts and no servers. The app doesn't send your data to us or to any other company. The partner game exchanges data only with the other device — over a Wi‑Fi network both devices share, or device to device when they are close together.
Where it is kept
Your settings, your guess sets — including the pictures you add — and your list of states are stored on your device, in the app's own storage. Until you turn on Save statistics on device, statistics are held only in memory while the app is running.
Like any other app's data, all of this goes into your device's backups.
The app contains no analytics and no crash reporting.
When you write to us
Email you send us is ordinary correspondence: we keep it only as long as it takes to answer you.
All the details are in the Privacy Policy.
Your own sets made with AI
You can ask an AI assistant, such as a chatbot, to draw pictures for a guess set, or to make a whole set. The pictures are SVG drawings. An SVG picture can contain its texts in several languages, and the app then shows the texts in the language the app is set to. A picture can also take the colours of the app's theme, in the theme's light version and in its dark version. A set can have its own “?” placeholder: the picture shown while the value is hidden. The assistant needs the description of the formats; each request below includes the address of this description.
Make a set
- Copy one of the requests below into the assistant and change it as you like: the topic, the number of pictures (from 2 to 200), the languages.
- If the assistant cannot open web pages, also give it the description of the formats: in Description of the Formats for AI, tap Copy and paste the description into the same chat, or tap Save and attach the saved file.
- The assistant answers with the set in one code block, often with a few words of explanation around it. Copy only the code block: in many chat apps it has its own copy button.
- In the app, tap on the main view, and on Add a Set tap Paste from Clipboard. If the assistant runs on another device, such as a Mac, save the code block there as a file with a name that ends in .json, and tap Import from file instead. The new set is added and chosen; tap Start Guessing to go to the main view. The same buttons are also right under these steps.
If there is an error, ask the assistant to fix it.
A set without captions:
Make a guess set named "Animals" for the app Intuition Check: 10 SVG pictures of different animals, without captions. First read the description of the formats and follow it: https://intuition-check.com/llm-formats.txtA set with a caption in each picture. The caption does not change with the language of the app:
Make a guess set named "Fruit" for the app Intuition Check: 8 SVG pictures of different fruit, each with a caption under the drawing: the name of the fruit. First read the description of the formats and follow it: https://intuition-check.com/llm-formats.txtA set with names in several languages. Each caption shows the picture's name in the language of the app:
Make a multilingual guess set named "Planets" for the app Intuition Check: 8 SVG pictures of the planets, each with a caption that shows the picture's name in the language of the app. Give the names in English, German and Hindi. First read the description of the formats and follow it: https://intuition-check.com/llm-formats.txtA set whose pictures take the colours of the app's theme:
Make a guess set named "Weather" for the app Intuition Check: 8 SVG pictures — sun, cloud, rain, snow, thunderstorm, wind, fog and rainbow — without captions. Their colours must follow the app's theme, in the light version and in the dark version. First read the description of the formats and follow it: https://intuition-check.com/llm-formats.txtMake one picture
The assistant can also make a single picture. Copy its code block, open one of your own picture sets, tap and tap Paste from Clipboard. The picture is added without a name: tap it and type one in Name. You can also save the code block as a file whose name ends in .svg, then tap in the set and choose Files; the file's name without .svg becomes the picture's name.
Draw one SVG picture of a cat for the app Intuition Check. First read the description of the formats and follow it: https://intuition-check.com/llm-formats.txtWhat else a set can have
Ask the assistant to explain what the formats allow, with an example request for each possibility:
Read the description of the formats for the app Intuition Check: https://intuition-check.com/llm-formats.txt Tell me briefly and in simple words what can be done in sets and pictures: how pictures can follow the app's theme — the different themes, the light and the dark version; how the names of a set and of its pictures can have translations into other languages; what kinds of captions a picture can have; what the set's “?” placeholder (its cover) can show. For each possibility, give an example request I can send you. Do not draw anything yet.Description of the Formats for AI
The description of the formats is written in English for an AI assistant, not for a person to read. The same description is on the website, at the address in the requests above.
This is a specification for language models; it is not formatted for human reading.
Format version: 1
Stable URL: https://intuition-check.com/llm-formats.txt
# Intuition Check: how to make SVG pictures and guess sets
Intuition Check is an iOS app. The app picks one picture of a guess set, hidden, and the user guesses which one.
A guess set is a named, ordered list of pictures. This text tells you how to produce (A) a whole guess set as
one JSON object, or (B) one picture as one SVG document, so that the app accepts it and draws it well.
The user's wishes always win. Every rule, default and style advice in this text is what you do when the user
has not said otherwise: anything the user asks for — any texts, any number of them, any place, size, shape,
colour or layout, a caption over the drawing, a word along an arc — overrides it. Only facts about the app
cannot be overridden, because the app enforces them (what it accepts and refuses, what it removes, how it
draws: sections 2, 3, 5 and the theme rules of 6 — but a sentence there that tells you what to do is still a
default, not a fact about the app); if a wish runs into one of them, do the closest thing the app allows and
say so.
## When the user asks what the formats allow (to explain, not to make)
Answer in the user's language, in plain words, without code. Keep apart what the app does and what you do by
default. The sizes, positions and orders in this text, and its sentences that tell you what to do ("put",
"give", "write", "keep"), are your defaults for a good result: the app does not check them and the user may
ask otherwise. Do not retell them as rules of the app ("must", "required", "not allowed"); say "by default I …".
What the app itself does:
- Themes: four themes, each light and dark. The app never recolours a picture on its own; a picture follows
the theme only where its own file says so (section 6), which you write when asked.
- Names: a set needs a name; a picture's name is optional. The set and every picture may have translations,
each its own languages; the app shows the one for the app language, else the default name.
- Captions: text in a picture is part of the drawing; the app never writes or translates it. A caption is one
fixed text, or holds versions in any number of languages, two included, and the app language picks one.
- The "?" placeholder: a set without a cover shows the app's own "?". The app draws no cover itself. A cover
is a picture like any other: it may show anything, with or without text, follow the theme and the app
language, and stand anywhere in the set.
- Limits: at least 2 pictures to guess (a cover is not one of them), at most 200 in all, 10 MB per SVG; one
refused picture refuses the whole set.
Do not translate the themes' code words (section 6).
Terms used below:
- name: the picture's name as stored in the app (`name`, `nameTranslations`). It is data, not part of the drawing.
- caption: a text drawn inside the picture (an SVG `<text>`). Optional. See section 4.
- app language: the language the app's interface is set to, in the app itself; it may differ from the device's
language. One of the codes in section 3.
- user's language: the language the user writes to you in, unless the user names another.
## 1. Output contract
A. The user asks for a SET (several pictures, "guess set", "set"):
- Answer with ONE JSON object as defined in section 2, in ONE code block. Nothing else is needed.
- If you add words outside the code block, they must not contain `{` or `}`: the app reads the pasted text
from the first `{` to the last `}`.
- Before or after the code block, tell the user in one short sentence, in the user's language, to copy the
code block (chat apps usually show a copy button on it) and paste it into Intuition Check as a new set — or,
on another device, to save it as a file whose name ends in .json and open that file in the app.
B. The user asks for a single PICTURE:
- Answer with ONE SVG document (section 5) in ONE code block.
- If you add words outside the code block, they must not contain `<svg` or `</svg>`: the app reads the
pasted text from the first `<svg` to the last `</svg>`.
- Before or after the code block, tell the user in one short sentence, in the user's language, to copy the
code block and paste it into one of their picture sets in Intuition Check; the picture is added without a
name. Or to save it as a file whose name ends in .svg and add that file to a set: the file name without
".svg" becomes the picture's name, so suggest a file name that is the name, e.g. `Cat.svg`.
In both: do not name the app's buttons or screens from memory, and do not translate their names yourself. The
app's help exists in every language of the app, with the exact names the app shows in that language; the steps
are in its section "#ai_sets", and the page for each language is listed in https://intuition-check.com/llms.txt. If
you can open web pages and the user wants the steps, take the names from the page in the user's language.
Otherwise the one sentence above is enough: the app's help and its own screens show the user where to paste.
Ask first, or proceed:
- Proceed without questions. Decide what the request leaves open, and say what you decided in a sentence or
two beside the code block:
- which things of a kind ("five animals": choose the animals);
- how many pictures, if no number is given: 4;
- the regional form of the user's language: take the likeliest and name it (e.g. "Canadian French");
- translations: as well as you can; name the languages you are least sure of;
- themes not mentioned: fixed colours, and one sentence that a version following the app's themes can be
made;
- a cover with nothing said about what it shows: the question mark (8c).
- Ask only when you cannot start: you cannot tell which language the names should be in, or the request names
no subject at all. Then at most 2 questions, in ONE message, in the user's language, each with the choice
you will make if the user does not mind; no code block in that message. Never ask what the message already
says. Plain words: no "code block", "JSON", "opaque", "plate", "import", nothing about how the format works.
- If you cannot ask (there is no dialogue), or the user tells you not to: proceed.
- What you say beside the code block is in the user's words: "its own background", not "opaque plate";
language names, not codes; at most four short points. The file endings .json and .svg are allowed: the
user needs them.
- Like all words outside the code block: no `{` or `}` next to a set, no `<svg` next to a picture. Do not
quote code outside the code block.
Defaults when the request does not say otherwise:
- No caption text in the drawing (section 4, option 1).
- No theme colours (section 6). Fixed colours of your own.
- Names: always give every picture a `name`, in the user's language. Give no `nameTranslations` unless the
user asks for other languages or for captions in several languages (section 4, option 4).
- Number of pictures: what the user asked for; never fewer than 2 or more than 200.
- No cover (section 2). A cover the user asks for is extra: it is not one of the pictures asked for.
## 2. The set format (JSON)
```json
{
"intuitionCheckSet": 1,
"name": "Animals",
"nameTranslations": { "ru": "Животные", "de": "Tiere" },
"images": [
{ "name": "Cat", "nameTranslations": { "ru": "Кот", "de": "Katze" }, "svg": "<svg xmlns=\"http://www.w3.org/2000/svg\" viewBox=\"0 0 100 100\"><circle cx=\"50\" cy=\"56\" r=\"30\" fill=\"#f57c00\"/><path d=\"M24 44 L28 14 L48 30 Z M76 44 L72 14 L52 30 Z\" fill=\"#f57c00\"/></svg>" },
{ "name": "Fish", "nameTranslations": { "ru": "Рыба", "de": "Fisch" }, "svg": "<svg xmlns=\"http://www.w3.org/2000/svg\" viewBox=\"0 0 100 100\"><ellipse cx=\"44\" cy=\"50\" rx=\"30\" ry=\"18\" fill=\"#1976d2\"/><path d=\"M70 50 L92 32 L92 68 Z\" fill=\"#1976d2\"/></svg>" }
]
}
```
(The drawings above are kept tiny to show the shape of the file. More examples are in section 8.)
Keys:
- `intuitionCheckSet` (required): the integer 1.
- `name` (required): the set's default name. Not empty.
- `nameTranslations` (optional, on the set and on each image): object, language code → name. Codes: section 3.
- `images` (required): array of at most 200 objects, at least 2 of them pictures to guess (a cover is one
more, not one of the 2).
Array order = display order. Each object:
- `name` (optional in the format; always give it): the picture's default name.
- `nameTranslations` (optional): as above.
- `svg` (required): one complete SVG document as a JSON string (section 5).
- `cover` (optional): the JSON literal `true` on the picture that is the set's cover (below). Leave it out
on every other picture.
- Unknown keys are ignored. Do not add any.
The cover: while the app hides the value to guess, it shows a "?" in its place. A set — any set, with or
without captions — may bring its own picture for that place: the cover. Give one ONLY when the user asks for
it (a cover, a back, a placeholder, a question-mark picture). It shows what the user wants: for a set of
playing cards, for example, the pattern of a card's back. If the user does not say what it shows, draw the set's
own frame, with a question mark where the other pictures have their subject (8c; the last picture of 8f);
when the pictures' plates differ in colour, a plate of the same family (8f: a neutral grey). Exactly one
picture gets `"cover": true` (if several do, the first counts and the others are ordinary pictures). The
cover is never an answer, so the set still needs at least 2 other pictures. Put it last in `images` (your default; the app takes it at any place). Give it
a `name` in words ("Question mark", not "?") and, in a multilingual set, `nameTranslations` in the same
languages as the pictures: the app shows the cover's name in the set's list of pictures, in the app language.
By default the cover has no caption (no word under it); the app draws any text it has. Its question mark is the cover's subject; it only borrows the
`class="caption"` pattern of section 4, option 4, so that it changes with the app language. Write the mark
as text (preferred) in two forms: `?`, and `؟` for Arabic and Urdu (the form those languages write). The app
does not change the mark itself: the cover holds both, and its own style sheet shows the one for the app
language. That works in every set, also in one without captions; the two marks are then its only `<text>`. Drawn as a
shape instead, the mark is the same in every app language: Arabic and Urdu see the Latin form. A cover with
a subject of its own (a card back, a cloud) needs no `<text>` and no style sheet.
A single picture (answer B) has no cover key: the user marks a picture as the cover inside the app. A single
picture meant as a cover and asked with a caption: do both, and say that the word will show on the cover.
Rules:
- Strict JSON, UTF-8: no comments, no trailing commas. Non-ASCII characters may be written as they are.
- Inside the `svg` string escape `"` as `\"` and `\` as `\\`; write the SVG on one line, or write each line
break as `\n`. No raw line break or tab inside a JSON string. Where you join two lines, put one space,
never nothing. No other escape is needed: an apostrophe stays `'` (`\'` is not JSON), and letters of any
script stay as they are.
- Types: `intuitionCheckSet` is the number 1, every name and translation is a string, `cover` is `true`. A
wrong type (`"cover": "true"`, a number as a name, `null` as a translation) makes the app refuse the whole
text.
- If the app refuses any one picture, it imports nothing and names the refused pictures. Every picture must
be valid.
- A name is one line of plain text. The app trims it, turns line breaks and tabs into spaces, and cuts it at
256 characters. Keep names short (one to three words): they are shown small.
How names are shown:
- A set is multilingual when the set or any image has at least one translation.
- In a multilingual set the app shows, for the app language: the translation with exactly that code; else
the translation with the code's first part (`fr-CA` → `fr`, `en-GB` → `en`, `es-419` → `es`); else the
default `name`.
- In a set without translations the app shows the default `name` in every app language.
When the user asks for a multilingual set:
- Default `name` (set and images): in the user's language.
- `nameTranslations`: one entry per requested language, on the set AND on every image. Give the same
languages on every image (the app does not require it: an image without an entry shows its default name).
- The language of the default `name` gets no entry of its own: a language without an entry shows the default
name.
- A regional code (`en-AU`, `en-CA`, `en-GB`, `es-419`, `fr-CA`) needs its own entry only when its word
differs from the `en` / `es` / `fr` entry. `pt-BR` and `pt-PT` have no common code: each requested one that
is not the language of the default `name` needs its own entry (asked for "Portuguese" with an English default
name: give both). The same for `zh-Hans` and `zh-Hant`.
- A regional form named alone ("British English", "Canadian French") goes under the parent code (`en`, `fr`,
`es`), in names as in captions: every form of that language then shows it.
- Languages not named ("a multilingual set"), or "all languages": every language of section 3 — 53 codes,
47 entries (see there). Translate as well as you can (section 1). Leave out only a language you cannot
write at all (it then shows the default name), and tell the user which.
- `nameTranslations` change only names, never a caption. A multilingual set has no captions unless the user asks for them; a
caption drawn inside the pictures needs its own words for each language in the SVG itself (section 4,
option 4).
## 3. Language codes
Write the codes exactly as here (case matters):
ar bg bn ca cs da de el en en-AU en-CA en-GB es es-419 fi fr fr-CA gu he hi hr hu id it ja kk kn ko lt ml mr ms
nb nl or pa pl pt-BR pt-PT ro ru sk sl sv ta te th tr uk ur vi zh-Hans zh-Hant
A code that is not in this list is not an error: the app drops it. An entry
with an empty value is dropped.
"All languages" is 53 codes but 47 entries besides the default, in names and in captions: the user's language
is the default and needs none, and `en-AU`, `en-CA`, `en-GB`, `es-419`, `fr-CA` show the `en` / `es` / `fr`
word unless theirs differs. (With a Bengali default, `en` is one of the 47 and `bn` is not.)
## 4. Captions (text inside the drawing): four options
Names (section 2) are separate from captions: the app never writes a name into a picture and never changes any
text of a picture. Give names always. Text in, on or under the picture — whatever the user calls it (name,
word, label) — is a caption request; a wish about the words alone (correct, careful) is not. If you leave out
something that was asked for, say so. Lettering that is part of the subject (STOP on a road sign) is
drawing, not a caption: allowed under every option, written once, with `lang`.
Urdu is drawn in Nastaliq, its own style, which reaches lower under the line than other scripts (about
0.3 × font-size): leave room under an Urdu caption. A `؟` in Urdu is smaller in Nastaliq: give it its own text and size (8c).
A caption's `lang` may be a regional code (`en-GB`, `fr-CA`), also where its name stands under the parent code.
Choose the caption option from the request:
1. DEFAULT — no caption. The drawing has no `<text>` that spells the name. Also for a multilingual set
whose user did not ask for captions.
2. ONE-LANGUAGE caption — when the user asks for a caption / label / word in or under the picture and names
no languages for it. A plain `<text>` with the word in the user's language; every app language shows
that same word. Put `lang="CODE"` of the caption's language on the `<text>`: a picture
with text is drawn anew for each app language, and without `lang` the letter forms follow the app language. No
`letter-spacing` in Arabic, Hebrew, Thai or an Indic script. Example:
`<text lang="ru" x="50" y="91" font-size="12" font-family="sans-serif" text-anchor="middle" fill="#222222">Кот</text>`
3. TWO LANGUAGES AT ONCE — when the user names exactly two languages for the word under the picture and does
not say it should change with the app's language. The same for any other text of the picture: each gets its
two lines. Two `<text>` lines, each with its `lang`, both always
shown, no style sheet for them: the user's language on top, font-size 10, baselines at y = 84 and 95,
subject in the upper 68 units (100-unit viewBox). A line in a script of 8b's "Lettering" block takes that
`font-family` as an attribute; a long word follows "Size" below. Names: the user's language as `name`,
the other in `nameTranslations`. Say that both words always show, and that captions that change with the
app's language can be made instead.
4. CAPTIONS IN SEVERAL LANGUAGES — when the user names three or more languages for the captions, or asks for
captions in all of the app's languages, or for captions that follow the app's language (= all of them).
The picture holds one caption per language and a small style sheet; through that style sheet the app
language picks the caption (8b, 8f). Languages: exactly those the user named; "all" = every language of
section 3 (53 codes, 47 entries: see there), translated as well as you can (section 1). In a set, give
every image `nameTranslations` with the same languages and the same words, so that the name and the
caption agree.
How option 4 works:
- A picture that contains `<text` or `:lang(` is drawn anew for each app language it is shown in, each time
inside an HTML page whose language is that app language: `<html lang="CODE">`, CODE from section 3 (`ru`,
`pt-BR`, `zh-Hant`, …). The app changes nothing in the file; the file's own CSS picks the caption with the
selector `:lang()`.
- `:lang(x)` matches a page in `x` and in every `x-…`: `:lang(fr)` also matches `fr-CA`, `:lang(es)` also
`es-419`, `:lang(en)` also `en-GB`. `pt-BR` and `pt-PT`, `zh-Hans` and `zh-Hant` are written separately.
- The pattern:
- All language versions of one text in ONE group with `class="caption"`; each `<text>` in it is one
language version, all at the same place.
- Any number of texts, anywhere and in any form (labels on parts, a title, words along an arc with
`<textPath href="#id">`): give each text its own group with `class="caption"` holding its language
versions; the one style sheet serves them all.
- The default caption has the class `default`. It carries the word of the default `name`, in the user's
language. It is shown when the page has no language and in every language that has no caption of its
own. The user's language is one of those: its code stands in no rule and on no caption. (In 8b and 8f
the default word is English, so `en` appears nowhere. The Zodiac files call this class `en`.) The
default caption is also shown on pages of other languages: give it `lang="CODE"` of the user's language.
If that language has a line in the "Lettering" block of 8b, add `.default` to that line's selector
(`.default, .hi, .mr { font-family: … }`; `.default` alone when no caption has that class), and to the
`letter-spacing: 0` line if you use letter-spacing.
- Every other caption's `class` lists the codes of the languages that show it. Languages with the same
word may share one caption: `class="bg ru"`; that changes only the `class`, never the style sheet. These
captions need no `lang`: each shows only on a page of its own language.
- The style sheet, in this order:
`.caption text { display: none }` — hide every caption;
`.caption .default { display: inline }` — show the default one;
`:lang(ar, bg, …) .default { display: none }` — hide the default one in every language that has its own
caption (list them all);
`:lang(ar) .ar, :lang(bg) .bg, … { display: inline }` — show each language's own caption.
- A regional code whose word differs from its parent's (`es-419` next to `es`): its own caption, its code in
both rules like any other, and one more rule: `:lang(es-419) .caption .es { display: none }` (the extra
`.caption` makes it win wherever it stands in the sheet). If its word is the parent's, give it nothing.
- A regional form named alone ("British English", "Canadian French") goes under the parent code (`en`,
`fr`, `es`), so every form of that language shows it. "Portuguese" alone: one caption
`class="pt-BR pt-PT"` when the word is the same.
- Do not repeat the default word under the user's own code (no caption, no rule, no `nameTranslations`
entry for it).
- Check: the codes in the hiding rule = the codes in the showing rule = the codes on the captions, and
each code is on exactly one `<text>`; the user's own code is in none of the three. Otherwise a language
shows no caption, or two at once.
- To leave a language out: no `<text>` for it and its code in neither rule. It then shows the default
caption.
- In a set: the same style sheet, the same attributes on the caption group and the same languages in every
picture, character for character. Only the words differ, and the size of a long word. (A cover has its
own text and rules: 8c, 8f.)
- Size: the app does not measure or shrink text. Choose one `font-size` on the group at which the captions
fit the free width. A word too long for it: first its own smaller `font-size` attribute, down to 9 (on a
100-unit viewBox); if it still does not fit, a shorter everyday word (and the same word as its name).
Never under 9. To estimate a width (measured in WebKit on a Mac):
- ordinary case in bold `sans-serif` (8f): 0.45 to 0.6 × font-size per letter. Треугольник is 77 of 100
units wide at font-size 12, and 68 at 10.5;
- a Chinese character: about 1 × font-size;
- Indic scripts, Thai, Arabic, Hebrew in bold `sans-serif`: 0.3 to 0.55 × font-size per code point (13
words measured).
- Case: ordinary case, as in 8f; capitals with wide spacing only for the style of 8b. Capitals follow each
language's rules: Turkish i → İ; Greek capitals lose the accent (ΑΣΤΈΡΙ → ΑΣΤΕΡΙ); German ß → SS.
- Fonts: copy the "Lettering" block of 8b (in a set, only the lines of the scripts you use). It gives every
script outside Latin, Cyrillic and Greek its own font list (fonts present on Apple devices, at a weight
they really have), and sets `letter-spacing` to 0 for Arabic, Hebrew, Thai and the Indic scripts — in most
of them the letters join or stack, and spacing breaks that (a line needed only when the group has
letter-spacing). Japanese, Chinese and Korean each get their
own font so that a Han character takes its regional form; that is why `zh-Hans` and `zh-Hant` keep
separate captions even with the same word.
- Direction: the page is left-to-right in every app language. Keep `text-anchor="middle"` for every script.
An Arabic, Hebrew or Urdu caption mixed with digits or Latin letters has not been tested: avoid it.
- A caption drawn twice (e.g. a shadow under it): put the group once inside `<defs>` with an `id`, and draw it
with two `<use href="#id">` that carry the colours and the offset.
- Whole set: captions and names in all languages fit in one answer. Write every language that was asked for, in
every picture. Do not ask about it, and do not tell the user it may not fit. In a set, leave out comments
and indentation.
- Never: fewer languages than asked without saying which are missing; captions left out when they were asked
for; one set spread over several messages or delivered as single pictures; a set cut short. A set is
always ONE JSON object in ONE message.
Start from 8f (a small set with captions and a cover, 100-unit viewBox). Take from 8b (a picture with a
caption in every language) its "Lettering" block, and its gold medallion style only when the user asks for
it; 8c is that set's cover. All three are complete and tested in the app. 8b and 8c are built like the app's
own Zodiac set; one of its files, if you can open web pages (you do not need it):
https://intuition-check.com/sets/zodiac/zodiac-08-scorpio.svg
## 5. The SVG document: what the app does with it
Facts (from the app's code):
- Size limit: 10,000,000 bytes per SVG (UTF-8). A larger SVG is refused.
- The app recognises an SVG by the characters `<svg` within the first 4096 bytes. Begin the document with
`<svg`: no XML declaration, DOCTYPE or comment before it.
- The app draws the SVG into a still bitmap, at most 1600 pixels on the longer side, with WebKit, inside an
HTML page. A picture with text is drawn anew for each app language it is shown in, and one with theme
rules for each theme. The background is transparent: where the drawing paints nothing, the app's own
background shows, light or dark depending on the user's theme.
- Scaling: the root's `viewBox` is fitted into the bitmap, aspect ratio kept, centred. `width` and `height`
on the root are ignored. A picture without a `viewBox` is accepted, but drawn unscaled (cut off or tiny):
always write one.
- The aspect ratio is read from the first `viewBox="minx miny width height"` in the first 4096 bytes: double
quotes, four numbers separated by spaces or commas, positive finite width and height. If it cannot
be read, the bitmap is square. Ratios outside 1:20 … 20:1 are clamped.
- `preserveAspectRatio="none"` on the root is replaced with `xMidYMid meet`: a picture is never stretched.
- Scripts never run. Nothing is loaded from the network: only content inside the file is drawn.
- Rendering has a time limit (10 seconds to load, 10 seconds to draw). A picture that exceeds it is refused.
Removed before drawing (so never write them; the rest of the picture is still drawn):
- these elements with everything inside: `script`, `foreignObject`, `meta`, `link`, `base`, `iframe`, `frame`,
`frameset`, `object`, `embed`, `applet`, `portal`, `form`, `input`, `button`, `textarea`, `select`, `option`,
`audio`, `video`, `source`, `track`, `template`, `noscript`;
- every attribute whose name starts with `on` (`onload`, `onclick`, …);
- every `href` / `xlink:href` whose value does not start with `#` (a part of the same file) or `data:image/`;
- `@import` in a style sheet; `<?xml-stylesheet …?>`; a DOCTYPE with an internal subset (`[ … ]`).
The removal works on the text of the file: the character sequences `on…=`, `href=` and `@import` are also
removed from text content and CSS, so do not write them there.
Allowed and drawn: shapes, `path`, `g`, `text`, `tspan`, gradients (`linearGradient`, `radialGradient`),
`pattern`, `mask`, `clipPath`, `filter`, `defs`, `symbol`, `use` with `href="#id"`, `<style>` with inline CSS,
`title`, `desc`, `image` with `href="data:image/…"` (a picture embedded in the file).
Not removed, but useless: `a` (a link does nothing in a picture in the app).
Rules for the SVG you write:
1. Root: `<svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 100 100">` … `</svg>`. Double-quoted
attributes. Lower-case `<svg`. One root; no nested `<svg>` elements.
2. Square `viewBox` (`0 0 100 100`) unless the user asks for another shape. The same `viewBox` for every
picture of a set.
3. Self-contained: no external images, fonts, style sheets or links. For text use
`font-family="sans-serif"` (or `serif`, `monospace`), or the font lists of 8b, which end with one of them:
the text is drawn with the device's own fonts.
4. Well-formed XML: every element closed, attribute values quoted, `&` `<` `>` in text and attribute values
written as `&` `<` `>`. Only SVG elements; no HTML elements.
5. Readable small: the app shows pictures in a grid, each cell may be small. One clear subject that fills
most of the viewBox; a few large shapes; strong contrast; no fine detail. On a 100-unit viewBox: the
subject's lines at least 2 wide, a caption's font-size at least 10, a long word 9. Ornament may be
thinner (the rings of 8b); in a small cell it may not be seen. The examples in this text are bare on
purpose; do not copy their simplicity. Build the subject from about 4 to 10 shapes: its outline, its
parts, and one or two details that make it this thing and no other. Someone who sees only the picture
must be able to name it.
Pale or pastel colours: keep the silhouette. The subject's main shape, or an outline around it, must differ
from the plate by at least 3:1 — a darker outline, or a deeper plate. That is for the silhouette only:
details inside the subject are free.
6. Distinct: the game needs the pictures told apart at a glance. Every picture of a set must differ clearly
from every other in silhouette or in main colour — not only in a detail. No two pictures with the same
SVG text.
7. One style across a set: the same viewBox, margins, line weights, palette family, and the same kind of
background (all with a plate, or all without). With captions: the same lettering and caption place in
every picture.
8. Background: the bitmap is transparent and the app's background can be light or dark. Either draw an
opaque plate of your own behind the subject (e.g. `<rect width="100" height="100" rx="12" fill="…"/>`), or
use only colours that stay visible on both white and near-black, or use theme colours (section 6).
9. Unique `id` values inside one SVG. Ids of different pictures do not clash (each picture is drawn alone).
Do not start an id with `ic-` unless it is a theme rule, and do not name a class of your own `ic-…` or a
variable `--ic-…` (section 6).
10. Keep it light: no heavy filters (large blurs), no deep `use` nesting, no animation (the app does not remove it, but
draws the picture as a still image).
11. A caption must not cross the drawing, unless the user allowed it (labels on the drawing's parts, a word on
a sign or over the subject are such wishes). Any text stays inside its own shape when it has one (a sign,
a label, a slice), and inside the picture. Keep the subject and the caption
apart — e.g. on a 100-unit viewBox: the subject above y = 78 and one caption line with its baseline at
y = 91; with two lines (section 4, option 3), the subject above y = 68.
## 6. Theme colours
Use this section ONLY when the user asks for pictures that follow the app's theme, or for light and dark
versions. Otherwise use fixed colours, no id that starts with `ic-`, no class `ic-…` and no variable
`--ic-…`. When the user said nothing about
themes: fixed colours, and offer a version that follows the themes (section 1). "Looks good in light and dark
mode" is not a theme request: an opaque plate with fixed colours does it (section 5, rule 8). A theme request
wins over a colour description in the same message ("gold on dark, matching the app's themes": theme rules).
The app has four themes — `sand`, `ocean`, `forest`, `dusk` — each in a `light` and a `dark` version. These words
(and `light`, `dark`) are code words for the rules below, not what the user sees: the app shows the four themes
under its own names in the user's language. When you talk to the user, do not translate the code words
yourself; say "the app's four themes" (and "light" / "dark" in the user's language), or, if you can open the
app's help page in the user's language (see section 1), take the themes' names from it.
Every version has its own values for five colours:
- `background` — the colour behind every view;
- `surface` — plates and panels on a view;
- `text` — the app's text colour;
- `accent` — the colour of buttons that are not filled;
- `soft` — in a light version the same as `accent` (a `soft` part on an `accent` part cannot be seen there);
in a dark version a pale tone close to the text colour.
The values (#rrggbb), in the order background, surface, text, accent, soft:
- sand light: f6f2ea eae3d5 403a35 645339 645339
- sand dark: 1b1a18 282622 c7c4bd b89b70 d2cbbe
- ocean light: eef4f7 dde8ee 363d40 3e5a6a 3e5a6a
- ocean dark: 0d1a23 16272f bec4c7 82a3b6 c6d4db
- forest light: f1f4ec e1e8d7 353e34 4a5e46 4a5e46
- forest dark: 111a14 1b261e bec4bd 8ba586 c9d4c6
- dusk light: f6f1f6 eae0ec 3f3944 5e4a6a 5e4a6a
- dusk dark: 17121c 231b2a c2bdc5 a894b4 d3c9d9
All four themes are muted: no colour above is bright. `text`, `accent` and `soft` on `background` or `surface`
have a contrast of at least 5.6:1 in every version.
Fixed colours and theme colours mix freely: give rules only to the parts that should follow the theme (a plate,
outlines, the main shape) and keep your own fixed colours where the picture needs them (the arcs of a rainbow,
a red apple). A part without a rule is drawn as written in every theme.
A part that keeps a fixed colour sits on a plate that changes: check it against the light and the dark
background and surface values of the table. Near-black vanishes on the dark ones (#1b1b1b: 1.01:1), white and
yellow on the light ones (#ffffff: 1.11:1, #f9a825: 1.54:1). Give such parts the theme's text colour — in a
style sheet `fill: var(--ic-text, #333333)` (or `stroke:`), with ids `ic-text` or `ic-stroke-text` — or
use a mid-tone (#e53935: 3.30:1 on the light ones, 3.57:1 on the dark ones).
Captions on a themed plate take the theme's text colour. In a style sheet: one more line,
`.caption { fill: var(--ic-text, #222222) }`, in the sheet that already picks the caption; a single `<text>`
gets a class and the same line; a caption drawn twice through `<use>` gets the class on the `<use>` that
carries its colour (8b). With ids: give the caption group `id="ic-text"` and a written `fill`; the group may
also carry `class="caption"`; a single `<text>` may carry the id itself. A themed cover's mark is coloured
the same way.
There are two ways to write theme colours, and the app reads both. Use a style sheet (6.1), also when the
user will edit the file in Inkscape (6.4). Use ids (6.2) instead when the user says the file will be edited
in Figma, Illustrator, Affinity or Sketch, or asks for ids: those editors write the file anew and drop its
`<style>` element, while ids can come through as layer names (Figma only with its option to include ids).
One way per picture, and in a set the same way in every picture: the app accepts both in one file, but a
rule in a group's id overrides the sheet colours of the classed shapes inside that group.
### 6.1 A style sheet
The app adds to the root `<svg>` the classes `ic-sand` | `ic-ocean` | `ic-forest` | `ic-dusk` and `ic-light` |
`ic-dark`, and defines the CSS variables `--ic-background`, `--ic-surface`, `--ic-text`, `--ic-accent`,
`--ic-soft` — but only if the file contains, in its `<style>` element, one of those six class names after a
dot (e.g. `.ic-dark`), or, in `<style>` or in a `style` attribute, a variable name that starts with `--ic-`.
```
.plate { fill: var(--ic-surface, #dddddd) }
.leaf { fill: var(--ic-accent, #e53935) }
.ic-dark .dot { fill: #ffd54f }
.ic-forest.ic-dark .dot { fill: #69f0ae }
```
The second value inside `var()` is the colour other viewers show. Give the styled elements a `class` and no
`fill` in a `style` attribute. The second value is needed: always write it.
- The five variable names are the only ones. Another name (`--ic-bg`) is never defined: its second value is
drawn in every theme.
- Write `var(--ic-…)` inside the `<style>` element, not in a `fill="…"` or `stroke="…"` attribute: the app
does not look there, and with no other mention it defines no variables.
- The app writes the six classes on the root. Never write one of them on an element yourself.
- A sheet rule beats a colour written in a `fill` / `stroke` attribute; a colour in a `style` attribute beats
the sheet.
- Of two sheet rules of equal weight the later one wins: plain rules first, then `.ic-light` / `.ic-dark`
rules, then theme rules.
- Put these lines into the same `<style>` as the caption rules of section 4, when the picture has them.
### 6.2 Rules in an id
A rule = optional limits + a colour.
- Limits, at most one of each kind, any order: a theme (`sand` | `ocean` | `forest` | `dusk`); a version
(`light` | `dark`); a paint (`fill` | `stroke`). No limits = every theme, both versions, fill and stroke.
- Colour: one of the five names above, or your own colour as exactly six hex digits (`1e88e5`). Colour words
(`red`) are not colours in a rule. All rule words are English.
`id="ic-` + one or more rules, all words joined by hyphens, own colours without `#`:
- `ic-text` — the text colour everywhere;
- `ic-light-accent-dark-soft` — accent in light versions, soft in dark versions;
- `ic-dark-1e3a2a` — own colour in dark versions; as drawn in light versions;
- `ic-ocean-dark-c3ced4` — own colour in the dark version of Ocean only;
- `ic-stroke-accent` — accent on the outline only.
- `ic-fill-accent-stroke-text` — two rules in one id: accent inside, text colour on the outline.
Reading stops at the first character that is not an ASCII letter, digit or hyphen, and at the first word that
is not a rule word. So keep ids unique with a numeric suffix: `ic-text-2`, `ic-text-3`. A word of your own
ends the rules: `ic-moon-text` holds no rule. A six-letter word of hex digits (`facade`) is read as a colour.
What a rule does:
- The id may be on a shape, a `text`, or a group `g`. A rule on a group covers everything inside it, except
what is inside `defs`, a gradient, `mask`, `pattern`, `clipPath`, `marker`, `symbol` or `filter`.
- For each element, fill and stroke are chosen separately. Its own rule is used when one applies in the
current theme and version; otherwise the rule of the innermost enclosing group that has one that applies;
otherwise the colour stays as drawn.
- Among several rules of one element that apply: theme + version beats theme, theme beats version, version
beats no limit; then a rule with `fill` / `stroke` beats one without; then the first one written.
- The chosen colour replaces a plain colour written in the element's `fill` / `stroke` attribute or in its
`style` attribute. Left as written: `none`, `transparent`, `currentColor`, `inherit`, `url(…)` (gradients),
`var(…)`, and colours with their own alpha (`#rrggbbaa`, `#rgba`, `rgba(…)`, `hsla(…)`).
- A `path`, `rect`, `circle`, `ellipse`, `polygon`, `polyline`, `text` or `use` with no `fill` written gets
the chosen fill added. A `line` does not. A stroke is never added: a rule colours a stroke only where a
`stroke` with a plain colour is written on that element itself (not only on its group).
- So, to be safe: write `fill="#rrggbb"` (and `stroke="#rrggbb"` where there is an outline) on every element
that a rule should colour. These written colours are also what other SVG viewers show.
- A `stop-color` or `color` written on a covered element is replaced too (by the fill colour, else by the
stroke colour). A gradient with a rule in its own id has its stops recoloured.
- Theme rules are not applied to an SVG with more than 50,000 tags; it is drawn as written.
- Do not colour the same element both ways. When an element with a `class` has a rule — its own or its
group's — the app writes the rule's colour into the element's `style` attribute, which beats the sheet
(unless the sheet says `!important`): a group's rule so overrides the sheet colours of the classed shapes
inside it. On an element without a `class` the sheet wins wherever one of its selectors reaches it.
### 6.3 Other ways the app reads (do not use unless asked)
- A `<desc>` directly inside an element, beginning with `Intuition Check`, then one rule per line as
`limits: colour` (e.g. `dark stroke: text`, `light: #1e88e5`); a first-line form `Intuition Check: text`.
- Inkscape swatches (a one-colour gradient whose label or id holds rules).
### 6.4 If the user wants to edit the file in Inkscape later
Inkscape keeps both ways, in Inkscape SVG and in Plain SVG.
- A style sheet (6.1): Inkscape keeps the `<style>` element with the `var()` inside it. Write `var()` only
there, never in a `style` attribute (Inkscape removes such an attribute when the element is edited), and add
`!important` to every colour in the sheet (Inkscape writes colours into the element's `style` attribute,
which otherwise wins).
- Ids (6.2): Inkscape keeps ids. Labels are lost in Plain SVG.
## 7. Checklist — run before answering
For a set:
1. The answer is one code block holding one JSON object; it parses as strict JSON.
2. `"intuitionCheckSet": 1` is present; `name` is not empty.
3. `images` has 2 to 200 entries: exactly the number of pictures the user asked for, plus the cover if
there is one.
4. Every image has a non-empty `name` and an `svg` string.
5. In every `svg` string each `"` is written `\"`, and there is no raw line break.
6. Language codes are from section 3, spelled exactly. In a multilingual set every image and the set have
the same languages.
7. Captions follow section 4: none by default; one language only if asked, with `lang` on the `<text>`
two named languages: two lines, both shown, each with `lang`; three or more, or all, only if the user
named them or asked for all. Then: the default caption has `lang`; the codes in the hiding rule = the
codes in the showing rule = the codes on the captions, each on exactly one `<text>`; the same style sheet
in every picture but the cover; every image has `nameTranslations` with the same languages and words.
8. A cover only if asked: one picture with `"cover": true`, last, plus at least 2 other pictures.
9. The answer is whole: the JSON ends with its `}`, and no picture was shortened or left for later.
10. Beside the code block: the sentence on how to use it, and what you assumed (section 1).
For every SVG (in a set or alone):
11. Starts with `<svg`, has `xmlns="http://www.w3.org/2000/svg"` and `viewBox="0 0 W H"` in double quotes,
ends with `</svg>`.
12. Well-formed: all elements closed, `&` `<` `>` escaped in text.
13. None of: `script`, `foreignObject`, `a`, `style` with `@import`, `on…` attributes, `href` to anything
but `#id` or `data:image/…`, external fonts, HTML elements, nested `svg`, animation.
14. Far under 10,000,000 bytes.
15. The subject is large, simple, high-contrast; the picture differs clearly from every other picture of the
set; it is visible on a light and on a dark background; no caption crosses the drawing.
16. Theme rules only if asked; without them no id starts with `ic-` and the style sheet has no `.ic-` class
and no `--ic-` variable. With them: one method in the picture, the same in every picture of the set.
Style sheet: `<style>` itself holds a `--ic-` variable or one of the six classes (a `var()` only in a
`fill` attribute does nothing); the variable names are the five of section 6; every `var()` has a second
value. Ids: every word after `ic-` is a rule word; the element has its `fill` (and `stroke`) written.
Every fixed colour left on a themed plate was checked against the light and the dark values.
17. Ids are unique within the SVG.
## 8. Complete examples
### 8a. Minimal set, no captions
```json
{
"intuitionCheckSet": 1,
"name": "Shapes",
"images": [
{ "name": "Circle", "svg": "<svg xmlns=\"http://www.w3.org/2000/svg\" viewBox=\"0 0 100 100\"><rect width=\"100\" height=\"100\" rx=\"12\" fill=\"#fff4d6\"/><circle cx=\"50\" cy=\"50\" r=\"30\" fill=\"#e53935\"/></svg>" },
{ "name": "Triangle", "svg": "<svg xmlns=\"http://www.w3.org/2000/svg\" viewBox=\"0 0 100 100\"><rect width=\"100\" height=\"100\" rx=\"12\" fill=\"#e3f2fd\"/><path d=\"M50 18 L84 78 L16 78 Z\" fill=\"#1565c0\"/></svg>" }
]
}
```
### 8b. One picture with a caption in every language (section 4, option 4)
A star in a gold medallion with its caption in all 53 languages, built like the app's Zodiac set: the
"Theme" block turns the gold to bronze in light versions (section 6.1), the "Language" block picks the
caption, the "Lettering" block gives each script its font. Languages with the same word share a caption;
es-419 and fr-CA follow es and fr; a page with no language, or one not listed, shows the default caption,
"STAR" (the user's language is English here). The "Theme" block is not part of option 4: this medallion has
no plate, and its gold has to darken on a light background. Without a request for theme colours leave that
block out and draw an opaque plate (section 5, rule 8), as 8f does.
```svg
<svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 512 512">
<style>
/* Theme. The app gives the root the classes ic-light or ic-dark (and ic-<theme>).
The gold is drawn for a dark background; on a light one it turns bronze. */
.ic-light .gold-top { stop-color: #A47A2A }
.ic-light .gold-bottom { stop-color: #6A4A14 }
.ic-light .gold-text { fill: #7A5A1E }
.ic-light .shadow-text { opacity: 0.12 }
.ic-light .shadow-glyph { stroke-opacity: 0.27 }
/* Language. Every caption is in the file once; its class names the languages that
show it. The page's language (html lang) picks the caption; without a language,
or with one not listed here, the default one. fr-CA follows fr, es-419 follows es. */
.caption text { display: none }
.caption .default { display: inline }
:lang(ar, bg, bn, ca, cs, da, de, el, es, fi, fr, gu, he, hi, hr, hu, id, it, ja, kk,
kn, ko, lt, ml, mr, ms, nb, nl, or, pa, pl, pt-BR, pt-PT, ro, ru, sk, sl, sv, ta,
te, th, tr, uk, ur, vi, zh-Hans, zh-Hant) .default { display: none }
:lang(ar) .ar, :lang(bg) .bg, :lang(bn) .bn, :lang(ca) .ca, :lang(cs) .cs, :lang(da) .da,
:lang(de) .de, :lang(el) .el, :lang(es) .es, :lang(fi) .fi, :lang(fr) .fr, :lang(gu) .gu,
:lang(he) .he, :lang(hi) .hi, :lang(hr) .hr, :lang(hu) .hu, :lang(id) .id, :lang(it) .it,
:lang(ja) .ja, :lang(kk) .kk, :lang(kn) .kn, :lang(ko) .ko, :lang(lt) .lt, :lang(ml) .ml,
:lang(mr) .mr, :lang(ms) .ms, :lang(nb) .nb, :lang(nl) .nl, :lang(or) .or, :lang(pa) .pa,
:lang(pl) .pl, :lang(pt-BR) .pt-BR, :lang(pt-PT) .pt-PT, :lang(ro) .ro, :lang(ru) .ru,
:lang(sk) .sk, :lang(sl) .sl, :lang(sv) .sv, :lang(ta) .ta, :lang(te) .te, :lang(th) .th,
:lang(tr) .tr, :lang(uk) .uk, :lang(ur) .ur, :lang(vi) .vi, :lang(zh-Hans) .zh-Hans,
:lang(zh-Hant) .zh-Hant
{ display: inline }
/* Lettering. A script whose letters join or stack gets no letter-spacing; every script
outside Latin, Cyrillic and Greek gets its own font. Telugu at 500: iOS has Kohinoor
Telugu only up to Medium, and WebKit would fake a 600 from it. */
.ar, .ur, .he, .bn, .gu, .hi, .mr, .kn, .ml, .or, .pa, .ta, .te, .th { letter-spacing: 0 }
.ar, .ur { font-family: system-ui, -apple-system, Geeza Pro, sans-serif }
.he { font-family: system-ui, -apple-system, Arial Hebrew, sans-serif }
.bn { font-family: Kohinoor Bangla, Bangla Sangam MN, sans-serif }
.gu { font-family: Kohinoor Gujarati, Gujarati Sangam MN, sans-serif }
.hi, .mr { font-family: Kohinoor Devanagari, Devanagari Sangam MN, sans-serif }
.kn { font-family: Noto Sans Kannada, Kannada Sangam MN, sans-serif }
.ml { font-family: Malayalam Sangam MN, Malayalam MN, sans-serif }
.or { font-family: Noto Sans Oriya, Oriya Sangam MN, sans-serif }
.pa { font-family: Mukta Mahee, Gurmukhi Sangam MN, sans-serif }
.ta { font-family: Tamil Sangam MN, InaiMathi, Tamil MN, sans-serif }
.te { font-family: Kohinoor Telugu, Telugu Sangam MN, sans-serif }
.th { font-family: Thonburi, Noto Sans Thai, sans-serif }
.ja { font-family: Hiragino Sans, Hiragino Kaku Gothic ProN, sans-serif }
.ko { font-family: Apple SD Gothic Neo, Noto Sans KR, sans-serif }
.zh-Hans { font-family: PingFang SC, Heiti SC, sans-serif }
.zh-Hant { font-family: PingFang TC, Heiti TC, sans-serif }
.te { font-weight: 500 }
</style>
<defs>
<linearGradient id="gold" gradientUnits="userSpaceOnUse" x1="120" y1="90" x2="400" y2="420">
<stop class="gold-top" offset="0" stop-color="#FBE9AE"/>
<stop class="gold-bottom" offset="1" stop-color="#C08F2C"/>
</linearGradient>
<!-- The captions. Drawn twice at the end of the file: the shadow, then the gold. A long
word would carry its own smaller font-size and letter-spacing. Urdu says lang="ur" and is
drawn in Nastaliq. zh-Hans and zh-Hant keep two captions
with the same word: each takes its own font. -->
<g id="caption" class="caption" text-anchor="middle" font-family="Helvetica Neue, Helvetica, -apple-system, sans-serif" font-weight="600" font-size="52" letter-spacing="6.8">
<text class="default" x="259.4" y="466">STAR</text>
<text class="ar" x="259.4" y="466">نجمة</text>
<text class="bg ru" x="259.4" y="466">ЗВЕЗДА</text>
<text class="bn" x="259.4" y="466">তারা</text>
<text class="ca es" x="259.4" y="466">ESTRELLA</text>
<text class="cs" x="259.4" y="466">HVĚZDA</text>
<text class="da nb" x="259.4" y="466">STJERNE</text>
<text class="de" x="259.4" y="466">STERN</text>
<text class="el" x="259.4" y="466">ΑΣΤΕΡΙ</text>
<text class="fi" x="259.4" y="466">TÄHTI</text>
<text class="fr" x="259.4" y="466">ÉTOILE</text>
<text class="gu" x="259.4" y="466">તારો</text>
<text class="he" x="259.4" y="466">כוכב</text>
<text class="hi mr" x="259.4" y="466">तारा</text>
<text class="hr" x="259.4" y="466">ZVIJEZDA</text>
<text class="hu" x="259.4" y="466">CSILLAG</text>
<text class="id ms" x="259.4" y="466">BINTANG</text>
<text class="it" x="259.4" y="466">STELLA</text>
<text class="ja" x="259.4" y="466">星</text>
<text class="kk" x="259.4" y="466">ЖҰЛДЫЗ</text>
<text class="kn" x="259.4" y="466">ನಕ್ಷತ್ರ</text>
<text class="ko" x="259.4" y="466">별</text>
<text class="lt" x="259.4" y="466">ŽVAIGŽDĖ</text>
<text class="ml" x="259.4" y="466">നക്ഷത്രം</text>
<text class="nl" x="259.4" y="466">STER</text>
<text class="or" x="259.4" y="466">ତାରା</text>
<text class="pa" x="259.4" y="466">ਤਾਰਾ</text>
<text class="pl" x="259.4" y="466">GWIAZDA</text>
<text class="pt-BR pt-PT" x="259.4" y="466">ESTRELA</text>
<text class="ro" x="259.4" y="466">STEA</text>
<text class="sk" x="259.4" y="466">HVIEZDA</text>
<text class="sl" x="259.4" y="466">ZVEZDA</text>
<text class="sv" x="259.4" y="466">STJÄRNA</text>
<text class="ta" x="259.4" y="466">நட்சத்திரம்</text>
<text class="te" x="259.4" y="466">నక్షత్రం</text>
<text class="th" x="259.4" y="466">ดาว</text>
<text class="tr" x="259.4" y="466">YILDIZ</text>
<text class="uk" x="259.4" y="466">ЗІРКА</text>
<text class="ur" lang="ur" x="259.4" y="466">ستارہ</text>
<text class="vi" x="259.4" y="466">NGÔI SAO</text>
<text class="zh-Hans" x="259.4" y="466">星星</text>
<text class="zh-Hant" x="259.4" y="466">星星</text>
</g>
</defs>
<circle cx="256" cy="210" r="148" fill="none" stroke="url(#gold)" stroke-width="6" opacity="0.85"/>
<circle cx="256" cy="210" r="136" fill="none" stroke="url(#gold)" stroke-width="2" opacity="0.5"/>
<path class="shadow-glyph" d="M256 112L279.5 179.6L351.1 181.1L294 224.4L314.8 292.9L256 252L197.2 292.9L218 224.4L160.9 181.1L232.5 179.6Z" fill="none" stroke="#0E0D22" stroke-width="19" stroke-linejoin="round" opacity="0.45" transform="translate(3,5)"/>
<path d="M256 112L279.5 179.6L351.1 181.1L294 224.4L314.8 292.9L256 252L197.2 292.9L218 224.4L160.9 181.1L232.5 179.6Z" fill="none" stroke="url(#gold)" stroke-width="19" stroke-linejoin="round"/>
<path d="M256 112L279.5 179.6L351.1 181.1L294 224.4L314.8 292.9L256 252L197.2 292.9L218 224.4L160.9 181.1L232.5 179.6Z" fill="none" stroke="#FFF6DA" stroke-width="4.5" stroke-linejoin="round" opacity="0.4" transform="translate(-2,-3)"/>
<g fill="none" stroke="url(#gold)" stroke-width="3" opacity="0.55" stroke-linecap="round">
<path d="M150 396h68M294 396h68"/>
</g>
<path d="M256 390 262 396 256 402 250 396Z" fill="url(#gold)" opacity="0.75"/>
<use href="#caption" class="shadow-text" fill="#0E0D22" opacity="0.45" transform="translate(2,4)"/>
<use href="#caption" class="gold-text" fill="#CA9E41"/>
</svg>
```
In a set this SVG is the image's `svg` string (on one line, `"` written `\"`). The image's `name` and
`nameTranslations` carry the same words as the captions, in ordinary case ("Star", "Звезда", "Étoile", …).
Every picture of the set lists the same languages.
### 8c. The cover of the set in 8b
The same medallion with a question mark instead of the subject (section 2, "The cover"); in a set it gets
`"cover": true`. Two marks: `?`, and `؟` for Arabic and Urdu. A cover may show anything else the user asks
for; this one is what to draw when the user does not say.
```svg
<svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 512 512">
<style>
/* Theme: the same rules as every picture of the set. */
.ic-light .gold-top { stop-color: #A47A2A }
.ic-light .gold-bottom { stop-color: #6A4A14 }
.ic-light .gold-text { fill: #7A5A1E }
.ic-light .shadow-text { opacity: 0.12 }
/* Language. Two marks: "?" for every language, the Arabic question mark for Arabic and Urdu. */
.caption text { display: none }
.caption .default { display: inline }
:lang(ar, ur) .default { display: none }
:lang(ar) .ar, :lang(ur) .ur
{ display: inline }
</style>
<defs>
<linearGradient id="gold" gradientUnits="userSpaceOnUse" x1="120" y1="90" x2="400" y2="420">
<stop class="gold-top" offset="0" stop-color="#FBE9AE"/>
<stop class="gold-bottom" offset="1" stop-color="#C08F2C"/>
</linearGradient>
<!-- The question mark, where the other pictures have their subject. Drawn twice below: the
shadow, then the gold. The marks differ in shape (Arabic ؟ in Naskh, Urdu ؟ in Nastaliq), so
each has its own size and place: the same ink box as "?". -->
<g id="mark" class="caption" text-anchor="middle" font-family="system-ui, -apple-system, sans-serif" font-weight="bold">
<text class="default" x="258" y="323.8" font-size="323">?</text>
<text class="ar" lang="ar" x="254.2" y="329.6" font-size="294.1">؟</text>
<text class="ur" lang="ur" x="258.4" y="325.6" font-size="317.3">؟</text>
</g>
</defs>
<circle cx="256" cy="210" r="148" fill="none" stroke="url(#gold)" stroke-width="6" opacity="0.85"/>
<circle cx="256" cy="210" r="136" fill="none" stroke="url(#gold)" stroke-width="2" opacity="0.5"/>
<use href="#mark" class="shadow-text" fill="#0E0D22" opacity="0.45" transform="translate(3,5)"/>
<use href="#mark" class="gold-text" fill="#E2C378"/>
<g fill="none" stroke="url(#gold)" stroke-width="3" opacity="0.55" stroke-linecap="round">
<path d="M150 396h68M294 396h68"/>
</g>
<path d="M256 390 262 396 256 402 250 396Z" fill="url(#gold)" opacity="0.75"/>
</svg>
```
### 8d. One SVG picture with theme colours (style sheet)
The plate takes the theme's surface colour, the moon the accent colour, soft in dark versions; the second
values are what other viewers show.
```svg
<svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 100 100">
<style>
.plate { fill: var(--ic-surface, #dddddd) }
.moon { fill: var(--ic-accent, #e53935) }
.ic-dark .moon { fill: var(--ic-soft, #e53935) }
.cut { fill: var(--ic-surface, #dddddd) }
</style>
<rect class="plate" width="100" height="100" rx="12"/>
<circle class="moon" cx="46" cy="48" r="28"/>
<circle class="cut" cx="59" cy="39" r="23"/>
</svg>
```
### 8e. One SVG picture with theme colours (ids)
The plate takes the theme's surface colour, the house the text colour, the door the accent colour; the sun is
accent in light versions and soft in dark versions. The written `fill` values are what other viewers show.
```svg
<svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 100 100">
<rect id="ic-surface" width="100" height="100" rx="12" fill="#eeeeee"/>
<g id="ic-text">
<path d="M16 52 L50 22 L84 52 Z" fill="#333333"/>
<rect x="26" y="52" width="48" height="30" fill="#333333"/>
</g>
<rect id="ic-accent" x="44" y="60" width="12" height="22" fill="#e53935"/>
<circle id="ic-light-accent-dark-soft" cx="82" cy="18" r="8" fill="#f9a825"/>
</svg>
```
### 8f. A small set with captions in three languages, and a cover
The user writes in English and asks for captions in English, German and Russian, and for a cover. The
default caption is the English word, so `en` is in no rule. Ordinary case, `sans-serif`, no letter-spacing;
the long Russian word has its own smaller size. Both pictures to guess have the same style sheet. The
cover's mark follows the app language, whatever the languages of the captions: `؟` in Arabic and Urdu, `?`
in every other. No theme rules: every picture has an opaque plate.
```json
{
"intuitionCheckSet": 1,
"name": "Simple shapes",
"nameTranslations": {"de": "Einfache Formen", "ru": "Простые фигуры"},
"images": [
{ "name": "Circle", "nameTranslations": {"de": "Kreis", "ru": "Круг"}, "svg": "<svg xmlns=\"http://www.w3.org/2000/svg\" viewBox=\"0 0 100 100\"><style>.caption text{display:none}.caption .default{display:inline}:lang(de,ru) .default{display:none}:lang(de) .de,:lang(ru) .ru{display:inline}</style><rect width=\"100\" height=\"100\" rx=\"12\" fill=\"#fff4d6\"/><circle cx=\"50\" cy=\"42\" r=\"28\" fill=\"#e53935\"/><g class=\"caption\" text-anchor=\"middle\" font-family=\"sans-serif\" font-weight=\"600\" font-size=\"12\" fill=\"#222222\"><text class=\"default\" x=\"50\" y=\"91\">Circle</text><text class=\"de\" x=\"50\" y=\"91\">Kreis</text><text class=\"ru\" x=\"50\" y=\"91\">Круг</text></g></svg>" },
{ "name": "Triangle", "nameTranslations": {"de": "Dreieck", "ru": "Треугольник"}, "svg": "<svg xmlns=\"http://www.w3.org/2000/svg\" viewBox=\"0 0 100 100\"><style>.caption text{display:none}.caption .default{display:inline}:lang(de,ru) .default{display:none}:lang(de) .de,:lang(ru) .ru{display:inline}</style><rect width=\"100\" height=\"100\" rx=\"12\" fill=\"#e3f2fd\"/><path d=\"M50 12 L82 70 L18 70 Z\" fill=\"#1565c0\"/><g class=\"caption\" text-anchor=\"middle\" font-family=\"sans-serif\" font-weight=\"600\" font-size=\"12\" fill=\"#222222\"><text class=\"default\" x=\"50\" y=\"91\">Triangle</text><text class=\"de\" x=\"50\" y=\"91\">Dreieck</text><text class=\"ru\" x=\"50\" y=\"91\" font-size=\"10.5\">Треугольник</text></g></svg>" },
{ "name": "Question mark", "nameTranslations": {"de": "Fragezeichen", "ru": "Знак вопроса"}, "cover": true, "svg": "<svg xmlns=\"http://www.w3.org/2000/svg\" viewBox=\"0 0 100 100\"><style>.caption text{display:none}.caption .default{display:inline}:lang(ar,ur) .default{display:none}:lang(ar) .ar,:lang(ur) .ur{display:inline}</style><rect width=\"100\" height=\"100\" rx=\"12\" fill=\"#eeeeee\"/><g class=\"caption\" text-anchor=\"middle\" font-family=\"system-ui, -apple-system, sans-serif\" font-weight=\"bold\" fill=\"#555555\"><text class=\"default\" x=\"50.4\" y=\"72.4\" font-size=\"64\">?</text><text class=\"ar\" lang=\"ar\" x=\"49.6\" y=\"73.7\" font-size=\"58\">؟</text><text class=\"ur\" lang=\"ur\" x=\"50.5\" y=\"72.8\" font-size=\"62.9\">؟</text></g></svg>" }
]
}
```
Contents
Short guides to the main things you can do in Intuition Check, and how to reach us.
- Intuition
- How the app works
- Guess sets
- Sending sets
- Partner game
- Statistics
- Reading your results
- Your state
- Appearance and features
- Free hour and subscription
- Your data
- Your own sets made with AI
- Version history
- Privacy Policy
If something isn't working
Write to us at [email protected] and we will answer. The quickest way is Settings → About → Contact support: it opens a draft in your mail app with the app version and build, your device model and your iOS version already filled in.
There is no support form and no ticket system. The app collects nothing, so anything we need to know has to come from you.
What to put in your message
- What you did, what you expected and what happened instead.
- The app version, from Settings → About → Version.
- Your device model and iOS version.
A draft from Contact support already has the version, the model and the iOS version.
Troubleshooting
- The two devices can't find each other: If the devices don't find each other.
- Statistics are gone after the app is reopened: Keeping your statistics.
- Nothing shows up in the statistics: What gets recorded.
- A set is missing from the switcher: The set switcher.
Sources
- Kahneman, D., & Klein, G. (2009). American Psychologist, 64(6), 515–526. Conditions for intuitive expertise: A failure to disagree
A joint review by two leading researchers of intuition from opposing schools, in the journal of the American Psychological Association.
https://doi.org/10.1037/a0016755 - Cardeña, E. (2018). American Psychologist, 73(5), 663–677. The experimental evidence for parapsychological phenomena: A review
A review of many studies by a psychologist who considers the evidence strong, in the journal of the American Psychological Association.
https://doi.org/10.1037/amp0000236 - Reber, A. S., & Alcock, J. E. (2020). American Psychologist, 75(3), 391–399. Searching for the impossible: Parapsychology's elusive quest
The reply to the review above by two psychologists who reject such effects, in the same journal.
https://doi.org/10.1037/amp0000486 - Sheldrake, R. (2000). Journal of the Society for Psychical Research, 64, 224–232. Telepathic telephone calls: Two surveys
Telephone surveys of randomly chosen households in London and Bury, England, by a researcher who holds that telepathy exists, in a journal of parapsychology.
https://www.sheldrake.org/files/pdfs/papers/Telepathic-Telephone-Calls-Two-Surveys.pdf - Brown, D. J., & Sheldrake, R. (2001). Journal of Parapsychology, 65, 145–156. The anticipation of telephone calls: A survey in California
A telephone survey of 200 randomly chosen people, by the same researcher and a co-author, in a journal of parapsychology.
https://www.sheldrake.org/files/pdfs/papers/The-Anticipation-of-Telephone-Calls-A-Survey-in-California.pdf - Cottrell, J. E., Winer, G. A., & Smith, M. C. (1996). Developmental Psychology, 32(1), 50–61. Beliefs of children and adults about feeling stares of unseen others
Three studies with children and adults, in a journal of the American Psychological Association.
https://doi.org/10.1037/0012-1649.32.1.50 - Born, J., Hansen, K., Marshall, L., Mölle, M., & Fehm, H. L. (1999). Nature, 397, 29–30. Timing the end of nocturnal sleep
An experiment with 15 volunteers, in the journal Nature.
https://doi.org/10.1038/16166 - Sio, U. N., & Ormerod, T. C. (2009). Psychological Bulletin, 135(1), 94–120. Does incubation enhance problem solving? A meta-analytic review
A review that combines the results of many experiments, in a journal of the American Psychological Association.
https://doi.org/10.1037/a0014212 - Wagner, U., Gais, S., Haider, H., Verleger, R., & Born, J. (2004). Nature, 427, 352–355. Sleep inspires insight
An experiment in the journal Nature; later repeats of it gave mixed results.
https://doi.org/10.1038/nature02223 - Ambady, N., & Rosenthal, R. (1993). Journal of Personality and Social Psychology, 64(3), 431–441. Half a minute: Predicting teacher evaluations from thin slices of nonverbal behavior and physical attractiveness
Three studies — the main one with 13 teachers — in a journal of the American Psychological Association.
https://doi.org/10.1037/0022-3514.64.3.431 - Diaconis, P., & Mosteller, F. (1989). Journal of the American Statistical Association, 84(408), 853–861. Methods for studying coincidences
A paper by two statisticians on how coincidences come about, in the journal of the American Statistical Association.
https://doi.org/10.1080/01621459.1989.10478847 - Wilson, T. D., Lisle, D. J., Schooler, J. W., Hodges, S. D., Klaaren, K. J., & LaFleur, S. J. (1993). Personality and Social Psychology Bulletin, 19(3), 331–339. Introspecting about reasons can reduce post-choice satisfaction
An experiment, in a journal of social psychology.
https://doi.org/10.1177/0146167293193010 - Dunn, B. D., Galton, H. C., Morgan, R., Evans, D., Oliver, C., Meyer, M., Cusack, R., Lawrence, A. D., & Dalgleish, T. (2010). Psychological Science, 21(12), 1835–1844. Listening to your heart: How interoception shapes emotion experience and intuitive decision making
Two experiments, in a journal of the Association for Psychological Science.
https://doi.org/10.1177/0956797610389191
Credits
Playing card artwork by Adrian Kennard, released under CC0 Public Domain. Generator and source at me.uk/cards.