Animai

Knowledge in motion, built in code

Comment

Inspiration

Great animated explanations make difficult ideas click (like 3Blue1Brown on Youtube!), but creating them usually takes specialized skills and hours of work. We wanted to make that process accessible to anyone, AND give people a platform to discover and share what they create.

What it does

The Animai engine is a code library that allows anyone to make a high quality animated educational video entirely in python code. It exposes functions with pre-built visual components, animations, color themes, voiceover, and a custom renderer, so you (or your agent!) just writes code for the video you want to make and it returns the rendered output video.

The Animai website is a social platform where users can create, share, and explore these animated educational videos. You can search for topics or scroll endlessly in a reels-style algorithm (but you're learning instead of rotting!). You can also create your own content with a topic, context, and style of your choice, which is fed into an agent which researches, synthesizes, and creates the video for everyone to watch.

How we built it

For the engine, we built a Python animation library on a WebGL-derived renderer, adding reusable educational components, visual themes, and synchronized voice narration. In the platform, Claude agents write and review the animation code, while Apify gathers relevant Youtube transcripts for research into other similar videos. FastAPI runs the generation pipeline, and Convex handles the data layer with accounts + realtime updates (eg. views and likes on videos).

Challenges we ran into

Making a high quality flexible animation library in code was challenging. Narration has to match the visuals on time, visual elements shouldn't be on top of each other even when moving, and all the styling should be consistent. We built a review loop into the platform's generation pipeline to ensure the results are more reliable.

Accomplishments that we're proud of

We connected a brand new code animation engine, an AI creation pipeline, and a social media style learning platform into one working product. Every generated lesson includes editable source code, giving creators a foundation they can understand, customize, and build upon.

What we learned

Especially in the AI era, it is extremely helpful to have strong code building blocks and to get feedback from actual renders outputs. In terms of the animation itself, we learned how much clear visual communication depends on timing, spacing, and consistency, and how reusable components help an agent produce better explanations.

What's next for Animai

Make the engine have even more components and customizability. For the platform, we can move rendering to scalable cloud workers and improve the discovery algorithm so users can enjoy browsing educational content naturally.

Built With

Share this project:

Updates

Submission history