MagicaLCore 2.0 goes from one model type to six: image, sound, tabular data, text, hand pose and hand action. Everything trains on the iPad's own chip, with no code, no Mac and no cloud. New in 2.0: pause a training and resume it later from History, test your model live with the camera or microphone, and a short explanation next to every setting. Export a standardmlmodel and drop it into Swift Playgrounds or Xcode. In 13 languages, plus an edition for schools.
Hi Product Hunt! I'm Damy, the independent developer behind MagicaLCore.
I wanted my iPad to be a real tool for building apps, not just a screen for watching things. So I made an app that trains machine learning models right on the iPad, with your own examples: one folder per class, and the folder name becomes the label.
With 2.0 you can train six kinds of models (image, sound, tabular data, text, hand pose and hand action), watch accuracy and loss as it learns, pause and pick it up later, and test it live with the camera or the microphone. When it's ready, you export a .mlmodel and use it in your own app with Swift Playgrounds or Xcode. All on iPad, with nothing uploaded anywhere.
It's free to start, with Pro if you need more. For schools there is MagicaLCore EDU, a one-time purchase edition: https://www.magicalcoreapp.com/edu
I'd love to hear what you would train with it, and what's missing. I'll be here all day answering questions.
No comment highlights available yet. Please check back later!
About MagicaLCore on Product Hunt
“Six kinds of machine learning models, trained on your iPad”
MagicaLCore launched on Product Hunt on November 3rd, 2026 and earned 73 upvotes and 1 comments, placing #27 on the daily leaderboard. MagicaLCore 2.0 goes from one model type to six: image, sound, tabular data, text, hand pose and hand action. Everything trains on the iPad's own chip, with no code, no Mac and no cloud. New in 2.0: pause a training and resume it later from History, test your model live with the camera or microphone, and a short explanation next to every setting. Export a standardmlmodel and drop it into Swift Playgrounds or Xcode. In 13 languages, plus an edition for schools.
MagicaLCore was featured in iPad (7.5k followers), Education (79.5k followers) and Developer Tools (522.4k followers) on Product Hunt. Together, these topics include over 134.9k products, making this a competitive space to launch in.
Who hunted MagicaLCore?
MagicaLCore was hunted by Damy. A “hunter” on Product Hunt is the community member who submits a product to the platform — uploading the images, the link, and tagging the makers behind it. Hunters typically write the first comment explaining why a product is worth attention, and their followers are notified the moment they post. Around 79% of featured launches on Product Hunt are self-hunted by their makers, but a well-known hunter still acts as a signal of quality to the rest of the community. See the full all-time top hunters leaderboard to discover who is shaping the Product Hunt ecosystem.
Want to see how MagicaLCore stacked up against nearby launches in real time? Check out the live launch dashboard for upvote speed charts, proximity comparisons, and more analytics.
Damy