Image: Confetti
Why it mattersA maker who cannot model a case or pick a connector can get a project from a conversation. The gate is whether the quality holds past a show-and-tell.
A person has an idea for a desk gadget, a kitchen tool or a custom controller. Between the idea and a real thing sit CAD modelling, parts sourcing, wiring and a trip to a print shop. Confetti wants to close that gap with one conversation. The company launched on Y Combinator as a Fall 2026 batch startup.
The founders are Tianyi Evans Gu and Lucas, both UPenn computer science dropouts. On their launch page they say they have won nine hackathons between them. They have done research at Stanford and Wolfram Research, and published at ICML. They built Confetti after a hard time making a home AI device of their own.
What Confetti does
The product has four parts, as the founders describe them on the launch page. Design generates and edits 3D models from a text or image prompt. Find parts sources off-the-shelf components. The same step helps with wiring and firmware for electronic projects. Get things made returns manufacturing quotes. It also sends supported 3D printing jobs to print farms, with parts delivered to the user. Share and remix opens the catalog of other makers' projects. A user can personalise a design for themselves.
The customer is a maker, not an industrial product team. The company lists worked examples on its launch page. One is a portable dynamic mechanical analyzer for a physics lab. Another is a custom keyboard built for faster typing. The list also names desk companions for an office and personalised gifts. Each of those is a one-off project, built once, sent to one person.
The YC traction and what is not public
Confetti had 24 votes on its Y Combinator launch page as of this reading. The company lists Fall 2026 as its batch and Consumer as its industry. The founders have not published a user count or a revenue figure. There is no lead investor and no stated round amount. Confetti.build is live, with an application wall that asks a visitor to apply for a workspace. Nothing else is public.
The founders' claim is bold. They write that AI has changed who can build software. They believe hardware is next. The test is whether a conversation really produces a project that works on the first print. That means the right parts in the box and firmware that boots. That is a different standard from a one-off demo. A buyer who waits for a part that does not fit will give up on the second try.
For a software engineer who has thought about building a device, Confetti is a short bet. The sign-up is free during preview. The output is a project file and a printed box. A failed first project costs the price of a few prints.
Source
- Confetti: Build hardware projects using plain-English prompts, Y Combinator
- Confetti, Confetti
This item was written by an AI system from the linked source. Reveneau is responsible for what it publishes.
Get AI News in your inbox
New developer tools, model and agent releases, and how teams are actually using them to release software. Short, and only when there is something worth reading.
