Velarium Studio

Designed for the worst day, not the best one

Most creative apps assume you've got time, taste and an idea. I made this one for the opposite. The person opening it might be awake at two in the morning with no energy and no clue what they want to make.

Everything else followed from that. If a screen asks you to decide something before it gives you anything, it has already lost the people I built it for. So the question was never what the app could do. It was how little it could ask of you and still be worth opening.

The practices come from published art therapy and music therapy writing, and the whole thing follows trauma-informed principles. It's made for entertainment and everyday wellbeing, not for treatment.

Rooms, not tools

It's built as places rather than tools. Drawing, painting, music, journaling, sensory work and sleep each get a room, with its own colour, its own guidance and its own idea of what finishing means. You go somewhere instead of picking a tool off a shelf.

The second level does more work than the first. Inside a room, sessions are grouped by how much they ask of you rather than by medium. A blank page and a page with something already on it are completely different experiences for the same person on different days, so I separated them and said so on the label.

Collapsing the browse list

At first every variation got its own card. The list got unreadable fast, and it was lying about the structure. Those weren't different exercises, they were the same session with a different starting point. Now one entry holds the whole family, and you choose the specific one inside the canvas after you've already started. You pick the kind of session first and the details second, which is the order people actually decide in.

A ceiling, never a filter

Once a day it can ask, optionally, how much you want the day to ask of you. Your answer caps what it suggests. Nothing gets hidden or locked, the whole library is still one tap away. It just aims lower. That's the difference between something that adapts to you and something that decides for you.

The scale everything is ordered by, and what the daily question caps.

A heart that's lit again

The star is doing the work. A heart with a notch in it is just a heart. Put a light in the notch and it turns into something that has come back on.

That's the whole promise of the app in one shape. Nobody arrives at an art therapy app feeling great. The idea was never that you get fixed, it's that something starts shining again, and the mark had to say that without being cheerful about it.

I drew the two halves wide and soft so they read closer to wings than to a valentine. The app is about letting your imagination off the leash as much as it is about feeling better, and a tidy symmetrical heart would only have said half of it. A cool rim light keeps the edges from going sugary, and a warm bloom sits behind the shape so it feels lit from somewhere rather than printed on.

The faint sparkles in the corners are the nostalgic part. Some of that is the dreamy, half-remembered quality I wanted the whole app to carry. The rest is that nostalgia feels like returning home, and that's the feeling I wanted someone to have when they open it. It's the same thing the tagline is after: a room you already know.

A palette that has to work at 3am

The direction is dreamcore. Dusk colours, soft grain, a bit of nostalgia. Lavender and peach for the daytime mood, deep violet and navy for the night one.

Pastels set a mood but they're useless as controls. Anything you confirm or destroy uses the deepest tone in the set, because a pastel button on a grey background is unreadable at exactly the moment someone needs to read it. Mood colour and interface colour are two different jobs.

Two axes that aren't the same axis

The app has its own day and night mood, and your phone has light and dark mode. They disagree all the time. Someone who keeps their phone dark permanently isn't asking for a dark app, they just want their eyes to be comfortable. I treated those as one setting at first, and it caused four separate bugs before I wrote the rule down: text that vanished, buttons you couldn't read, a room that quietly stopped matching the rest.

Everything shares the same paper

One rule holds six differently coloured rooms together. Every room sits on the same ground, and nothing paints over it. Rooms tint, they don't replace. It sounds like nothing, and it's the main reason the app feels like one place instead of six.

Three roles. A bookish serif for the voice, a neutral sans for the interface, monospace wherever numbers change while you watch them.

The interface as a kind of care

Rotation speed is a safety decision

One room has kaleidoscopes that turn slowly. The obvious question is how fast. The one that actually mattered was different. A turning radial figure doesn't flicker at the speed it rotates, it flickers at that speed times the number of repeating segments. Twelve segments turning once a minute flickers twelve times a minute, which is not what "once a minute" sounds like.

Flicker in a certain range is a real seizure risk, so the speeds sit a long way under it, with plenty of room rather than just enough. The limit gets checked automatically instead of left to my judgement. Switch on Reduce Motion and they hold one still frame, because a kaleidoscope that has stopped is still nice to look at.

The daily question asks about the work, not about you

It asks what you want the day to ask of you. An earlier version asked how much you had in you and I cut it, because capacity for what? You can have plenty to put on a page and nothing left for anything outside it. Same words, opposite answers. So the question points at the session instead of the person.

It never tells you things are getting worse

After a session you can say whether you're leaving lighter, heavier or about the same. It will never tell you the pattern is going badly. When the answers lean heavy it offers gentler work and says nothing about the tally, because reflecting that back feels like being watched. Things often get worse before they get better and the timescales vary wildly, so every reading backs you for showing up rather than for how it went. Tests enforce that, not good intentions.

Defaults that know what you came for

Whatever tool you're holding when a session opens depends on what kind of session it is. Something broad for filling in colour, something fine for detail work. Small thing. It takes away one decision at the point where decisions cost the most.

Every tool says what it does to the page, not just what it's called. You don't have to already know what a sponge brush is to pick one.

What testing changed

I tested with real people and changed the app based on what they told me. Two of those conversations turned into features.

A blank page can be the hardest part

One person told me that free drawing was intimidating on bad days, and that sitting in front of an empty canvas took more out of them than it gave back. What helped was being told what to draw. They weren't asking for less creative freedom. Finding the idea was the expensive part, and on a hard day they didn't have it to spend.

That became the guided half of the drawing room. The app brings something to the page so you don't have to turn up with an idea. You still make every mark, you just don't have to invent the starting point. It's the first thing I point people to now when they say they don't know what to do.

Some frequencies were making things worse

Someone using the body scans in the sleep section told me parts of the audio felt uncomfortable. The low end was dropping further than it should and it was causing them distress, which is about the opposite of what a body scan is for.

So I set limits at both ends, high and low, and went back through the whole library to check every piece against them instead of only fixing the one that got reported. Comfort is a measured property of the sound now, not something I assumed was fine.

A constraint that turned into the identity

There are no sound files in the app, no video, and no saved images of the artwork it gives you. Every sound is made while it plays and every visual is drawn as you look at it. It started as a size constraint and became the character of the thing. Sessions never repeat, artwork is sharp at any size, and the whole app stays small enough to download on a bad connection.

Generating the audio also forced a discipline I'd have skipped with files. Loudness is matched by measurement across the library so moving between pieces never jumps, and the inaudible low rumble is filtered out because it was eating headroom and doing nothing.

Left: a drawing session, guidance on top and a deliberately small toolbar underneath. Right: the music room, the one place I broke my own palette rule.

The music room is dark in both moods. An instrument wants a stage, and all that softness made the keys harder to find. Breaking your own rule once, on purpose and for a reason you can say out loud, isn't the same as not having one.

Subtracting in public

The first release shipped a set of visual exercises that never worked. They were waiting on content that never made it into the app, so all they did was show a loading spinner forever. The second release deleted them and put something generated in their place, which can't fail that way.

The release notes say all of this plainly, including an apology for a crash that took me too long to find. An app for people who have been let down by things that promised to help can't be cagey about its own mistakes. Writing that down was as much a design decision as any colour was.

What this is, and what it isn't

• It's a real app. Live on the App Store for iPhone and iPad, currently on its second release.
• It's trauma-informed, and built on art therapy principles taken from published writing on art therapy, music therapy and somatic practice.
• It isn't therapy. Not treatment, not a medical device, and no substitute for proper care. It's made for entertainment and everyday wellbeing.