Lucid

Inspiration

Online classes make it very easy for a student to fall behind without anyone noticing.

A professor can explain a difficult formula or concept, move on, and the student may simply stay quiet because they don't want to interrupt the class.

We wanted to solve a simple problem: what if students could see the explanation instead of only hearing it?

That idea became Lucid.

What it does

Lucid turns lectures into visual learning.

It takes a lecture transcript and identifies the important concepts. Then, instead of giving students more paragraphs of text, it creates the best visual explanation for each concept — such as a diagram, illustration, formula breakdown, or key points.

The goal is simple:

Listen → See → Understand.

How we built it

We built Lucid around an AI pipeline that takes lecture content, breaks it into smaller concepts, and decides how each concept should be explained visually.

For every concept, the system can generate a different type of visual depending on what makes the most sense.

We also created a concept thread that organizes the important topics from the lecture, so students can quickly find and revisit something they didn't understand.

The current version works with transcripts, but the same pipeline can eventually connect to real-time speech-to-text for live classes.

Challenges we ran into

The biggest challenge was figuring out how to visualize a concept correctly.

Not every topic should be shown as the same type of image. A physics formula might need a breakdown, while a process might need a diagram, and another concept might be easier to understand through an illustration.

We also had to make sure the visuals were actually useful for learning and not just visually attractive.

That pushed us to focus on one question:

“What is the clearest way to explain this concept?”

Accomplishments that we're proud of

We're proud that Lucid can take something as simple as a lecture transcript and turn it into a visual learning experience.

The part we're most proud of is the concept thread.

Instead of giving students one long lecture transcript, Lucid creates a visual map of the important concepts and lets students quickly go back to the topic they missed.

We also created the foundation for taking Lucid from a transcript-based tool toward a real-time classroom experience.

What we learned

We learned that AI doesn't always need to give students more information.

Sometimes the problem is not the amount of information — it's how that information is presented.

A student may understand something immediately when they see a diagram that they couldn't understand from a paragraph of explanation.

We also learned that good AI products need to focus on the user's actual problem, not just what the technology can generate.

What's next for Lucid

The next step is to make Lucid work during a live lecture.

Instead of waiting for a transcript, Lucid could listen through real-time speech-to-text and build the visual board while the professor is teaching.

We also want to add more subjects, improve the quality of generated visuals, and allow students to ask questions about any concept directly from the visual.

Our long-term goal is simple:

Turn every lecture into something students can see, understand, and follow — in real time.

Built With

• artificialintelligence
• computervision
• generativeai
• machine-learning
• naturallanguageprocessing