zu teuer!
viel zu teuer! schade weil an sich eine gute idee. 350€ nur für aufschlag ist zu viel. das muss auch günstiger gehen vielleicht mit werbung ansonsten buche ich mir einfach einen trainer und hab für weniger geld mehr davon.
300$ per year paywall
This buggy app which doesn’t work is behind a 300$ dollar per year subscription. Insanely steep price for something that doesn’t help your game at all. Much less worth 30 dollars a month. Feels like a rush job. There are definitely better alternatives to this.
The worse part is the AI ‘support’
Uploaded a video of my serve to the app, and the feedback I got was ‘no mistakes’.
The support team consists of an AI robot waffling on about general issues. It’s told me to add better prompts ‘comparing it to top pros’, (fairly sure there is no option to add prompts) ‘take a side shot for more accurate analysis’ (a direct contradiction to the instructions). ‘The app is probably analysing general flow and not biomechanics’.
The app is also really clunky and slow.
Don’t trust these slick YouTube videos. So far it’s been a complete waste of money.
The support team consists of an AI robot waffling on about general issues. It’s told me to add better prompts ‘comparing it to top pros’, (fairly sure there is no option to add prompts) ‘take a side shot for more accurate analysis’ (a direct contradiction to the instructions). ‘The app is probably analysing general flow and not biomechanics’.
The app is also really clunky and slow.
Don’t trust these slick YouTube videos. So far it’s been a complete waste of money.
Show more
Great idea, lacking execution so far
I echo the other review, though maybe it has gotten better by now with a larger dataset to work from. When I used it in Oct 25, it was not able to accurately distinguish different parts of my body and joint angles even if I set up everything perfectly with good lighting and distinguishable clothing. I thought I was going crazy trying everything to change my stroke until I realized it was not reading my stroke right in the first place. Additionally, although Mark is a good researcher and can identify problems with technique, he isn’t an expert coach in fixing those problems, so the tips for improvement are quite limited.
Show more
underdeveloped
his product feels completely raw and underdeveloped — hard to believe they’re charging a premium for it. The so-called “AI” serve evaluation gives ridiculous and inaccurate feedback; I even tested it by recording a player who didn’t serve at all, and it still produced detailed “issues” in every stage of the serve. It’s a great idea in theory, but in its current state, it simply doesn’t work. Definitely not worth $60 a month — and by the way, the video processing alone takes a good 15 minutes.
Show more
Response from developer
Hey thanks for the feedback. So Bert was training on videos of people serving, and it assumes that when someone uploads a video, there's someone serving in it. So when you uploaded a video with no actual serve, Bert was still looking for all the stuff we training it to look for - that's why you got a long list of mistakes. Bert is basically saying, "Hey I didn't see you do XYZ in this video." In terms of the processing time, we now have it down to under a minute :-) Thanks again for the feedback. Helps us improve Bert. Will
