Tiny update. I’ve worked on turning all sheets, like the rule one, into full-screen pages. A component I noticed in some apps that I love. This enables a smooth transition, sliding a sheet up and closing it by dragging or pulling it out.
This is an under-the-hood update, but much needed since I released Zooom.
With the release of the Fable model, I thought it’d be interesting to ask it to look into performance improvements.
Zooom used to re-encode the whole video from scratch, just to attach audio. Not anymore. The export now copies the already-encoded video straight through instead. Also, the Photo mode used to redraw the full-resolution photo into a fresh buffer every single time, 450 times. This made for very slow processing. Now it draws it once and reuses that same buffer 450 times.
Lastly, the library grid used to refetch your whole zooom album every time you opened it or recorded something new. It now listens for changes instead, and thumbnails are cached rather than decoded fresh on every scroll.
I’m working on a fourth kind of rule: time limits.
I might release it end of August, early September.
You can set the number of minutes you want to spend on one or many websites/apps, and when they’re spent, the shield comes down until the next day.
One tricky thing that I’ve yet to figure out is how to reliably count minutes with the iOS Screen Time integration. It’s limited.
Thirty seconds turned out to be shorter than kids’ thoughts. It takes a good 10s, at least for my kids, to figure out what to say after they press the big red record button. So I’ve extended the recording to a minute. So far it’s spot on.
To make it more obvious, I’ve also turned the room number sitting in the navigation header into a red countdown. When it hits zero, the clip sends as usual.
In mhm, the emoji is as important as the audio you send. It flavours what you say.
Now the emoji picker lets you choose from all 1,953 of them. A search field on top helps you find the right one for your mood.
mhm makes sounds now. I wanted the sounds to match the buttons’ look and feel. For example, the mic button clicks in and out. I used the Epidemic Sound library over the ElevenLabs one as they felt more “clicky” and less “generated by AI”.
New rooms don’t start empty anymore. There’s a clip already waiting, a short 👋 from me explaining the basics.
I used ElevenLabs voice generation to have a friendly voice say something like “Hey, you’ve just joined a room, everything is tappable. Look at that big red button…”.
I always wanted to find meaningful quotes to ask users to confirm, words that wouldn’t feel empty. I made up a couple of quotes while working on the rest of the app, and now it feels like they need attention.
The quotes are now borrowed from authors I respect, around the topic of attention, silence and the present moment. Words from Woolf, Thoreau, Pascal, Weil, Marcus Aurelius, Dillard, and Rilke. Each one is credited under the line as you type.
I made a new app. It’s called mhm.
I built it for my kids, who are 3 and 7, as we’re travelling separately for two weeks. It’s a voice-messaging app that lets you record short audio messages and send them to a room.
The week before leaving, I was anticipating wanting to talk directly to my kids, and I was curious to see if I could build it in 3 days. Also curious to see if and how they would actually use it.
It’s super simple and relies only on a single button to exchange messages. First you enter a room. A room is a three-digit code and a password (set by me for now).
You land on a single screen split into two sections. The top one has a big red button that lets you record your message and pick an emoji to attach to it, for a bit of extra fun.
The bottom section is for recorded messages, yours and theirs, one play/pause button on each.
It’s all designed to be used by a 3+ year old. You press to record your message and after 30 seconds, it’s sent automatically. Everyone in the room gets a notification when a new mhm arrives.
For now, room creation is closed as I’m working on the beta version. I don’t know if this will turn into a public tool.
Now when a new mhm lands in the room you’re logged into, you receive a notification with the attached emoji.
A bell button in the nav lets you turn those notifications on or off (off by default).
I received an email from Benjamin, a recent user of Cactus, asking for “an option for a stricter mode”.
So I’ve added a toggle called Set in stone. I might rename it, but it essentially removes the ability to delete a rule.
I’ve added some words of caution when it’s turned on, so you know that enabling this mode stops you from editing, deleting or removing any apps/websites from the rule.
The only way out is to delete Cactus.
One thing I found frustrating was not being able to quickly add a website to be blocked by a rule. You’re browsing and figure out in the moment that this website is sucking way too much of your attention.
Instead of having to open Cactus, find the rule, add the website and save the changes, you can add a website to a rule straight from Safari.
I’ve started to build a new mode. I haven’t found the right name for it yet, something like “Emergency mode” or “Moment”. I’m unsure.
I built it so you can block almost everything except a few apps you choose ahead of time, from a single tap. Sometimes you need a quick break from everything, and it’s the perfect feature for that.
I’ve been dogfooding Cactus quite a bit, and I noticed that taking the first break feels different from taking the second, third, and so on.
The first break is sometimes just a quick unlock to check on something important, so I’ve adjusted the friction around rule breaking a bit. The first break is friction-free, the second asks you for a short quote confirmation, and the following ones all ask for a longer one.
The logic resets the next day at midnight.
My friend Nicolas is building a sleep habits app and integrated some of the Screen Time features. For the blocking overlay, he used a “frosted” glass effect.
I thought that was neat and implemented a similar one for Cactus. Now you can see the app or website you’re blocking right there behind the shield.
I was going on a trip to the Adirondacks and wanted to shut off some distractions only for that trip. So I built the ability to set an end date on a rule.
When the trip or limit is over, the rule turns inactive and the blocking stops.
Rules used to only allow scheduled hours and days. Now you can set them to always on and block distractions without a break.
I’ve changed the way the first rule is created a bit. I started with a regular rule sheet, the same one we use for all rules.
When I tested Cactus with less-savvy users, they found it overwhelming.
With the latest update, the first rule creation is a completely different flow. Each step has a dedicated screen, and I’ve added a bit of storytelling around it to teach users what to expect and how it works.
Among the things that I discovered while working with Apple’s Screen Time feature, the most annoying one is that you can’t block an app or website for under 15 minutes.
You can choose from three icons for Cactus. My favourite is the original one, the Spark. Eventually, I want to create a suite of icons. That could be really fun.
Now scheduled rule cards count up or down depending on their status, with a small ticking animation.
Zooom got its spring cleaning: a new icon, simpler screens, and a proper pass for iOS 26.
I redesigned the library sheet, cut down the buttons, and made recording glow around the edges of the screen so you always know when it’s on.
The onboarding flow is way better too. There’s a welcome screen now, and the app asks for permissions properly instead of assuming you’ll figure it out.
Cactus is a pretty simple app, but some concepts need an explanation. My take for Cactus was to avoid or delay a help section as long as possible and rely on educating users after positive moments, a rule created, a break stopped…
The screen time app I’ve been prototyping has a name now: Cactus.
I hesitated between Zen, Quiet and Calm, but I found those names a bit corny. While I was hesitating between those, I thought I’d try to create the logo for the app and landed on a spark.
A spark that resembles the shock you might experience when touching something you shouldn’t, a cactus or a screen time rule.
I’ve fixed a major loophole in the app. If you were to reboot your phone, all rules and shields would be cleared, and unless you opened Cactus, nothing would be blocked.
Now Cactus checks on launch whether a rule should be active and puts the shields back itself.
Zooom is now bidirectional. You share a picture from anywhere via the share sheet and turn it into a zooom with your favourite music loop.
Everything you record with Zooom lands in its own album, and now you can browse it without leaving the app.
The library button wears the thumbnail of your latest video. Tap it and you get the grid: everything you’ve made, with durations, a share sheet, and a little haptic tick on every touch.
Zooom now takes photos. Sort of.
Tap instead of drag, and the app captures a single photo, then turns it into a 15-second video with your selected loop playing over it. A still image with a soundtrack, ready to share anywhere.
Getting there was a battle: upside-down photos, mirrored videos, audio that stopped halfway. A full day of fixes for a feature that looks like nothing. That’s usually a good sign.
I’ve published a tiny site presenting the Beta version Zooom. It’s simple, but playful, just like Zooom is so far.
The beta is now open to the public, and I got a couple of strangers (who knows how they found the link) and friends/family members testing the app. Try it for yourself
You can zoom in two directions, from both front and back cameras, with 4/5 loops. I’m adding more loops and features slowly.
This is my to-do list at the time of writing. I’m struggling to add the halo shown on the video preview in the export (and ideally as an export option). The next most important item is the onboarding. I’m not super excited to work on this, I don’t know why, but I have to do it sooner than later.
Filming yourself zooming in on your own face with a sad saxophone underneath is exactly as ridiculous as it sounds.
About a month ago, I set out to prototype the revival of an app I loved—one that existed back in the day, back when Instagram had 8 photo filters and no reels.
My main trade is software design, not engineering. I can hard-code HTML/CSS, I understand the basics of React but I know nothing, absolutely nothing about Swift.
For this app, I wanted to build it on my own—for two simple reasons: the complexity of the App seemed relatively low, and I wanted to move quickly while I was still excited about it.
I found in the past that, yes, you can go further with the help of friends or others, but the effort to coordinate and/or agree on the smallest details takes the fun away from these types of side projects.
For this one, fun is the main goal. So over the past few weeks, I’ve spent a couple of hours here and there, dabbling in what the used-to-be cool kids call vibe coding.
A couple of prompts to Cursor, and the app was running locally on my phone. I specifically skipped the
For this V1, I just wanted the app to run. Run smoothly, so I could show it as quickly as possible. Right now users can pick between two audio loops and record a video from the back camera.
The recorded video only includes the audio from the loop. If I remember correctly, that wasn’t possible with the original app, maybe due to some technical limitations.
I showed and shared the prototype with a couple of friends—some familiar with the original, some not. So far, the feedback and process have been nothing but positive. We got some good laughs out of it.
In 2017, the New York-based studio This Also released an iOS app called OOO.
OOO let you zoom or un-zoom while recording a video and overlay it with looped music. It had 5 to 6 different “themes” and was super simple to use. Press and move up or down on your screen to record and control the zoom, then release to stop and save the video.
Two examples — the sound is ON:
The first and main reason is selfish: I want to use the app. I’ve been wanting to use it periodically since it was taken off the App Store.
Another reason is that music filters as good as Zooom’s now exist solely on social media apps. Unless you’re recording and publishing through those apps, you don’t get the fun filter.
Start with one music track — likely the same “Sad Saxo” from the examples above — one zoom direction, all while recording a video.
The video should be stored in the user’s photo library, ideally under a separate folder. The interface can be minimal, but we’d need permissions to record and save videos.
I’m going to start building the first version in the next few weeks. I’m not even sure I’ll end up publishing it, and if I do, it’ll probably be free or cost less than $2.
For the last few weeks, I’ve been trying different apps that would turn my phone into something closer to the minimalist Light Phone.
I tried a few: Blank, Dumb Phone, and Smile App Launcher—the latter being the one I stuck with the most. Yet, despite their best efforts, launching an app wasn’t the smoothest experience.
So I went back to a more integrated and simple solution using the Shortcuts app and custom icons.
I only kept my most important apps: Camera, Homecam, Safari and Messages (Clockwise).
I also kept the minimal and uniform initial design, but to differentiate each app, I chose shapes that I associate with each one.
It’s been a week since I made the change, and I’m still happy with it. If you have any questions, feel free to reach out—I’d be happy to help.
Manuel Moreale recently proposed the concept of hiding specific topics.
A similar idea would let users define personal keywords to blur on the browser level. I would, totally randomly, choose to blur terms like AI, Trump, Growth, you get the idea.
Rather than completely removing content, the browser extension would apply a CSS blur filter to the parent block containing specified keywords, making them less visually intrusive.
In what could only be described as a shameless love letter to Things’ iconic ”+ New List” button (hey, if you’re going to steal, steal from the best), we added two fresh menu items to Type.
One for importing notes from other apps, because apparently some people write outside of Type.
Type’s screen real estate is limited. So naturally, we added a contextual menu for managing notes, rather than cluttering each note item with options.
And yes, keyboard shortcuts included, because we’re not monsters.
After experimenting with a pay-to-try model, Nicolas was exploring a new paywall design for Type. He had the essentials in place - features, pricing, and basic structure - but wanted to test different approaches.
Nicolas had nailed the essential content structure with clear feature descriptions and pricing. But we saw an opportunity to make it more compelling with some thoughtful tweaks.
We improved the grouping, added user reviews, refined the visual hierarchy, and introduced a lifetime pricing option.
Make sure to give a try to Type if you haven’t already.
In late 2022, Nicolas was really getting into one of his side projects, Type, a note-taking app for the busy mind. He reached out for some help, and designing an app icon was a fun and different challenge from my day-to-day work. It was something I hadn’t done in a long time.
The process was relatively short. Here are most of the steps that led to the final icon.
Starting with quick sketches helped us align on the different directions we liked and disliked. Without much doubt, the idea of a keyboard key made more sense than any other idea.
The first few high-fidelity keys had a retro feel but were kind of bland. I tried a couple of alternatives with tints or combination of more than one key.
A key by itself wasn’t strong enough, but surrounding the main key with other keys made it feel more “realistic”. Keeping the focus on the main key “T” was our final decision.
