A DIY door opener blew every fuse in Samuel Beek's house. He rebuilt the instructions — with Anthropic's Claude.

From a blown fuse to a new tool

Samuel Beek started Schematik after a project went badly. He'd followed wiring advice generated by a chatbot to build an electric door opener and discovered the bot had confused wet and dry connections. The result was a surge that knocked out the house wiring. Beek admits he isn't a hardware person, but the mistake pushed him to build software that understands hardware better than the general-purpose chatbots he first tried.

He moved his prompts from a different chatbot to Claude, the large language model from Anthropic, and layered a set of controls and checks on top. The resulting assistant — Schematik — takes a plain-English idea for a device and turns it into a shopping list, wiring diagrams and step-by-step assembly guidance.

What Schematik does

Beek says Schematik should speed up making gadgets the same way developer tools speed up writing software, by handling repetitive bits like parts selection. Tell it what you want — an audio player, a simple home automation trigger, a Tamagotchi-style toy — and it will propose components, show how they connect and link to places you can buy the parts. It also walks users through construction, offering practical instructions rather than abstract code snippets.

Beek describes the project as "Cursor for Hardware." He pitches it as a sidekick for makers — it tackles the boring stuff like sourcing parts and wiring so people get to testing sooner. For people who can solder but dislike sourcing parts, Schematik is meant to shorten the time between idea and finished device.

Early users and projects

The service drew attention after Beek shared it on X in February. Hobbyists and tinkerers tried it out and posted their results. Marc Vermeeren, who leads branding at N8N, used Schematik to put together several small devices. He built an MP3 player and a Tamagotchi-like bot called Clawy that helped him manage Claude coding sessions.

Other builders took the idea further, creating playful variants — one Clawy took on a character look inspired by television — and reported that Schematik's shopping and wiring suggestions reduced the usual back-and-forth of DIY projects. People posting finished builds on X and forums drew others in and boosted early curiosity.

Funding and commercial plans

Beek is preparing to turn Schematik into a business. The project has already attracted outside capital: venture capital firm Lightspeed Venture Partners invested US$4.6 million. Beek says he plans to monetise the tool eventually, though he has kept the early iterations available for people to use right away.

Lightspeed's $4.6m check shows investors think there's money in tools that turn prompts into real-world parts and circuits. Schematik's model — combining a large language model with domain-specific safeguards and shopping integrations — is one clear way to package that idea for consumers and makers.

Mastheads

Safety and accuracy challenges

The opening example — a miswired door opener that wiped out fuses — shows the limits of generic chatbots when they advise on physical systems.

Electrical work, mechanical tolerances and component compatibility all carry real risk if instructions are wrong or incomplete. Beek's switch to Claude, and the subsequent engineering of Schematik, reflect an effort to bake in the right kinds of checks and more specialised knowledge into the assistant.

DIY projects also break when people pick the wrong parts, ignore power needs, or build things that don't fit together. Schematik's pitch is to surface those details automatically: casting a wide net over the parts a project needs, then narrowing to the correct versions and safety-oriented wiring instructions.

How users report their experience

Users who tried Schematik described a faster build cycle. Vermeeren said he completed several projects with the assistant's help. For some makers, the biggest time-saver was the parts list and direct links to purchase the right components, rather than spending hours hunting for compatible modules and connectors.

That process matters for both beginners and experienced builders. Beginners get a guided path that can reduce early failures. Experienced makers get a quicker route from idea to prototype. Community members who've shared their builds online have also produced forks and stylistic tweaks to the core projects, showing the tool can serve as a platform for iteration.

Technical and business trade-offs

Linking hardware workflows to an LLM creates two clear headaches. First: how to keep the model current with part availability and price changes. Schematik provides buy-links for parts, but keeping those links accurate will need ongoing work and data sources. Second: quality control. A suggestion that a component will work in a circuit must reflect real electrical specifications — voltage, current, tolerances — not just approximate compatibility.

Beek says he rebuilt his workflow on Claude to make the assistant more practical — in other words, he forced it to give usable parts lists and assembly steps, not just abstract ideas. Those extra layers are likely where Schematik will differentiate itself commercially: users will pay for reliability and for the minutes saved avoiding a bad build.

Where this fits in the maker ecosystem

Tools that automate parts of the maker process aren't new. Online communities, parts marketplaces and CAD platforms have long reduced friction for physical projects. Schematik combines those ingredients with a conversational interface, and that might make the path from idea to device feel more natural to non-experts.

For professional engineers, Schematik probably won't replace detailed design tools or lab testing. For hobbyists, educators and rapid prototypers, an assistant that stitches shopping, wiring and assembly into a single flow could shrink the barrier to entry for hands-on projects.

The social angle

Part of Schematik's momentum came from people showing off finished devices. That social sharing matters: when a maker posts a working build, other people can replicate or adapt it. The ability to export a reproducible parts list and instructions makes for easy sharing, and the examples posted by users act as both marketing and informal validation.

Beek's early posts and the projects from people like Vermeeren created a loop: more projects generate more interest, which brings in more testers, which yields more projects.

Related Articles

"That's the difference: Your fuses blow out, or you have a solid product," Samuel Beek said.

This article was created with AI assistance.