Lamina Labs Is Building Visual Infrastructure for EdTech AI – Turning Text Prompts Into Accurate Explanations in Seconds

Language models have proven remarkably good at explaining things in words. The trouble is that learning is not a purely verbal act. Diagrams, animations, and visual intuition do more of the work of education than any textbook sentence. Lamina Labs, a Y Combinator Spring 2026 startup, is building an AI explainer video generator that closes that gap – turning a student’s text prompt into an accurate whiteboard animation in seconds.

“We are building the fastest visual infrastructure for EdTech AI applications,” the company says. “With a simple text prompt, students get accurate visual explanations in seconds, making high-quality learning dramatically faster and cheaper.” The bet is that the next wave of education technology will not be defined by chat interfaces alone but by systems that can generate – on demand and at scale – the kind of figures, animations, and worked examples that currently require graphic designers, subject-matter experts, and production pipelines to produce.

Lamina Labs is led by Kartikesh Mishra and Sudip Rokaya, co-founders with deep MIT roots. Mishra earned his BS in 2024 and MEng in 2025, both at MIT. Rokaya is on leave from MIT’s CS and Math program to build the company. Both founders are explicit that the product is infrastructure rather than a consumer app: Lamina Labs wants to be the rendering and reasoning layer that tutoring apps, adaptive learning platforms, homework tools, and creators of instructional content can build on top of – the way modern apps build on top of payment, messaging, or map APIs.

The market opportunity is broad and growing. The global EdTech market is measured in the hundreds of billions, and the most successful AI-native entrants – from homework helpers to corporate learning platforms – are racing to find the right medium of instruction. Text alone leaves too much on the table; video is expensive to produce at the pace of individualized learning. A reliable API that can turn a student’s question into a clear visual explanation would unlock a new class of products, in much the same way that hosted transcription and translation unlocked entire categories of consumer audio tools.

Technically, the challenge Lamina Labs is tackling is difficult. Generating accurate educational visuals requires more than stylish graphics; it requires understanding the concept behind a question, choosing the right representation – a diagram, a graph, a step-by-step animation – and getting the underlying reasoning right. The founders’ MIT computer science and mathematics backgrounds show through in this choice of problem. Speed and accuracy are both non-negotiable: students won’t wait, and teachers won’t tolerate visuals that subtly mislead.

With a lean two-person team and a Y Combinator Spring 2026 cohort debut, Lamina Labs is positioning itself as a picks-and-shovels play for the AI-native education wave. The company’s ambition is to make the economics of high-quality visual learning materials – once reserved for the best-funded EdTech brands – accessible to anyone building in the space. If Lamina Labs can deliver on the promise of accurate, on-demand whiteboard animations, Simi could become the default AI explainer video generator that the next generation of EdTech products is built on.