sinaatalay

Generative UI for education

https://academa.ai

Thank you very much! That’s strange, we’ll look into it. Really appreciate you reporting it.

The lecture description language isn’t public at the moment. It’s a core part of the technology we’re building the company around.

Also, if you’d like, you can share the video as a public link by clicking the share button!

We think videos are special. There is a reason why Khan Academy, Coursera, etc. are essentially video libraries. A lecture video is a recording of a teaching performance, a form of presentation where someone explains while controlling what you see and when you see it.

Also, you’re not chatting with the transcript; you’re chatting with the video. The AI has much more than the transcript in its context. See this, for example: https://academa.ai/lectures/diffusion-models-learning-to-den...

And in the coming months, we’ll make the answers themselves real-time videos, which is kind of trivial for us at this point. We’ll ask the LLM to answer by writing code, and our software will render that code and show you the response directly as a video.

We ask Claude Opus 5 to write a specific kind of source code that only our proprietary software understands and can render into a lecture video with computer graphics and TTS.

Without our system, there isn’t really an equivalent target language you can ask the LLM to write. Manim exists, but you won’t get videos like these by simply asking an LLM to generate Manim code.

Thank you!

Actually, the economics are quite good.

We’re not using video generation models. We ask an LLM to write the lecture as code, then render it deterministically with computer graphics and TTS.

So the main costs are the LLM call, TTS, and some cloud GPU time for rendering. That ends up being much cheaper than generating long form video directly with video models.

You can attempt to do it, but you'll see that it won't work. We've built the underlying "lectures as code" technology to make this work.

Claude Opus 5, with high thinking, generates the video in a single shot.

The one that asks questions at the beginning is Gemini Flash 3.7.

On consumer devices, AI communicates with us through speakers and screens. Screens are the richer medium, so most consumer AI innovation will happen there.

Computer graphics will have enormous applications because they are directly controllable by LLM-generated code. Video models are probabilistic and less suitable when precision matters. In education, for example, we need exact visuals. If an AI wants to plot y = sin(x), it should generate the precise graph through computer graphics rather than approximate it with a video model.

Sorry about that paywall. We've made the AI chat free up to a certain usage limit, so if you're interested, you can give it a try now!

Thank you for letting us know about the Safari issue.

We've made the AI chat free up to a certain usage limit, so if you're interested, you can give it a try now!

Thank you for letting us know.

Thank you :)

We are fully compatible with Manim CE. Our goal is 1:1 API compatibility, so people can continue using the API they already know.

We never intended to use 3Blue1Brown's name in a negative or misleading way. We simply wanted to communicate that we have made the Manim API created by 3Blue1Brown accessible in the browser with WebGPU, without requiring any installation.

You can use the editor however you like without signing up. The examples are simply sample prompts for the AI, but you can paste in any Manim code to editor from ChatGPT or anywhere else, and the preview will render it instantly.

Downloading videos is completely free. Only the AI-powered features are behind the paywall, and there's a 7-day free trial.

We'll put more pre-written examples, thank you!

Thank you!

Thank you! We simply mount the entire app inside the landing page, but rotate and scale it down. When the page opens, it smoothly animates into full screen using CSS transitions.

Let me elaborate.

Our content is text, not MP4 files. So we review it with humans and AI, just like how software companies review their software. We may make mistakes, but that's okay because we maintain the content. We continuously improve it and fix issues as we find them.

It'll even become more accurate than cam-recorded lecture videos over time, because those aren't being updated.

Thank you!

Thank you for the feedback! We'll make it better. We take quality very seriously.

Everything I do is driven by a single goal: to make my startup successful so I can innovate in edtech. Nothing else.

You may not believe me, and you may think I have ulterior motives, but I don't. My only goal is to build technology that meaningfully advances education.

First, I need to make my startup successful so I can continue innovating over the next 15 years. Open-sourcing our work could be a distraction at this stage. Maintaining an open-source project requires significant time and effort.

It would also make it easier for competitors to build products that compete directly with us, such as https://studio.academa.ai. That would make it harder for my startup to succeed.

I'm not building this for the sake of making money. I simply need enough protected space to innovate competitively and give the company the best chance to succeed. If it does, I'll be able to keep building and pushing education forward for years to come.

These are 14 hours of video, so it's easy to find mistakes. We've probably made plenty of them. But this isn't AI slop.

We created the technology so that the content can be updated and improved over time. If we make mistakes, we'll fix them and continue maintaining the material. These aren't static MP4 files recorded once with a camcorder.

Thank you for the feedback!

We're an edtech startup, Academa: https://academa.ai

We're two PhD students and close friends since our undergraduate years.

I'm Sina: https://sinaatalay.com

And my co-founder, Apo: https://geduk.io

All we want to do is innovate. We simply can't open source this because we're rewriting and rebuilding it every day.

There are some incorrect premises here.

"The videos are LLM slop." - No, they're not. Watch them.

"It's dogshit that no one can understand." - That's not true. Try it.

I couldn't agree more. But if you look at it from this perspective, you may understand why we use static videos for this demo.

Definition: the AI tutor is some AI that can generate these videos.

If that generation happens in real time, it's a complete reimagination of education. It's literally the AI private tutor.

But before you can do that in real time, ask yourself: let alone real time, can you even generate a video, let's say, in 30 minutes? That's what we explored, and we found some great things. But it's not real time yet. The only practical way to publish what we'd built, though, was as a static video paired with chat.

The complete reimagination of education is on the way.

No, not yet. It's something we may do in the future.

We are. We're confident in the quality of the teaching and the accuracy of the content in these videos.

Our videos are not some MP4 files sitting on a disk. We see our content like software. We review it, version-control it, and maintain it.

We have an in-house computer graphics pipeline. We reimplemented 3Blue1Brown's Manim in Rust from scratch. It is available at https://studio.academa.ai. The renderer runs in the browser using WebGPU.