How SmashSpeed built a speed tracker used by 50,000 players

It's an iOS and Android app where you give it a video and it measures the speed of your smash using computer vision and physics. It aims to bridge the gap between technology and badminton, which is usually a niche sport with much less accessible technology compared to bigger racket sports like tennis or padel.

In badminton, and a lot of other sports, there isn't really an accessible ranking system. I wanted to create a benchmark for players to compare with each other. Right now the only quantifiable metric is something like provincial ranking, and that isn't accessible because you have to play a lot of tournaments to earn points. I wanted one accessible metric to compare players.

I got the idea in early 2025. I knew there would be computer vision and machine learning involved and I didn't think I was ready, since grade 10 school coursework is pretty light. So I did two to three months of coursework on Coursera learning machine learning and the prerequisite math. At the beginning of the summer I actually started building. I pretty much camped at a café for around 12 hours a day and got the app done. After that came the hard part, marketing. I thought at first there wouldn't be much demand, but we went viral pretty quickly, I saw how easily you can get traction, and everything grew exponentially from there.

At first we did social media on Instagram and TikTok, and those went well. After that it was one user telling a friend, and that friend telling another friend, so we got a lot of free marketing from word of mouth.

In the results screen we explicitly made shareable graphics that you can export, because we wanted to make it really easy to share with your friends. I also think the idea itself is viral enough that you want to tell a friend.

First you pick a language. The two biggest options are Swift, which is only for Apple, or React Native, which works for both Apple and Android. We started on Swift but quickly realized React Native was much more efficient. After you code the app you pay for an Apple developer account, about 99 USD a year. Then you fill out a lot of metadata like descriptions, support URLs, and marketing URLs, and sign compliance contracts and policies for each country you want to distribute to. Those steps are mechanical and intuitive. You submit to Apple and they usually give a few rounds of feedback on specific details that get flagged, for example needing a delete-account feature or not having a hard paywall. Once you fix those, it is straightforward and gets accepted.

The best strategy was finding a viral format that was easy enough to keep reproducing and also feasible to make different each time so it doesn't get repetitive. For us that was the interview format: we ask someone how fast they think they can smash, show their actual smash, then show their reaction. It works because it goes viral and it's easy to get a lot of them. You head out to a tournament or your local badminton club, film a bunch in an hour that you dedicate purely to filming, and then you have about two weeks of content to edit.

When we post reels we don't expect to go viral every time. We expect the bare minimum, maybe one viral post for every 10 or 15 reels, because our reels are really reproducible and not that high effort. We can flush out around 10 reels, and as long as one goes viral we get a lot of new users. We don't need every one to go viral.

First, the idea itself has to be very creative and niche. If the idea is not viral, then no matter how much you market it, it's a waste of time. For coming up with ideas, you basically stare at a wall and think about what problems exist in your circle and what creative solutions could fix them. It sounds general but that's literally what you do. If you don't have experience, do a few weeks or months of coursework getting the foundations of the coding language you're using and any math it requires. There are a lot of coding agents that write the math and code for you, but you should still understand to some extent what your code does. For marketing, get a group of friends, because it's more fun and marketing is all about creativity, so the more brains you have the more chances of a creative idea.

Your public image for whatever app you're creating is very important, and that starts with your logo. Your logo has to have a very clear color theme. It could be a stencil simple enough that someone could draw it out. That extends to your landing page, your social media page, and your username. All of it sounds trivial, but it's someone's first impression of your app. Even if your idea is not that good, if the image is good, people will be more lenient about trying it out.

If you want to learn from SmashSpeed, follow us on Instagram at smashspeedai, and on TikTok, LinkedIn, X, and YouTube. You can also check out smashspeed.ca, and the app is on the App Store and Google Play.