Schematik Is ‘Cursor for Hardware.’ Anthropic Wants In

Samuel Beek knew he had a problem when he blew every fuse in his house. The culprit was an electric door opener he had built himself, guided by instructions for wiring and piecing together a device drummed up by ChatGPT. Turns out, the chatbot wasn’t so great at distinguishing between wet and dry connections, so the device he had built sent out a surge of misallocated power that zapped everything else. Oops.

Beek, based in Amsterdam, admits he is not a hardware guy. But he had that itch and now really just wanted to make something that wouldn’t explode. He switched his requests to Anthropic’s Claude, then rejiggered that into an assistant program he calls Schematik and has described, over and over again, as “Cursor for Hardware.”

The idea of Schematik is essentially vibe coding for physical devices. Tell the program what you want to make, and it will suggest just about everything you need to build it out in the real world. Beek is working on integrating a shopping list, so you can buy the individual wires and pieces. Then, it will serve as a guide for putting it all together. Beek plans to make money off it eventually and is working on getting investors. (It just got $4.6 million from venture capitalist firm Lightspeed Venture Partners.)

When Beek posted on X about the idea in February, it got lots of traction. Other tinkerers gave it a shot, describing what they wanted to make and then building it out. On Thursday, Anthropic engineer Felix Rieseberg posted on X to announce that Anthropic has now enabled “a little Bluetooth API for makers and developers, allowing you to build hardware devices that interact with Claude.”

“The big problem in hardware is that it's very gatekept and that very few people can do it,” Beek says. “I really hope that my tool can help enable more people to build, either with the tool or to learn how to build hardware through the tool.”

Kyle Wiens, CEO of iFixit, says he hasn’t used Schematik but thinks it could be useful. He notes that electronics design can be very complex, often requiring sorting through many different SKUs and ensuring compatibility of all the pieces. “That is just a super hard problem,” Wiens says. “This kind of scale is the sort of thing that AIs are good at.”