ClearClass

My daughter sent me the timetable she'd received from school and asked me to print it.

The original was almost comically difficult to read.

Tiny text, too much information squeezed into each lesson, and a layout that occupied two pages when one should have been enough.

I took it to Claude and transformed it into a colorful, readable, single-page timetable.

I shared the result with the other parents. For a day, I was something of a hero.

Then I thought: why shouldn't every parent be able to do this?

The concept was simple. Upload whatever timetable the school provided and receive something you can actually use.

I decided to turn it into a publicly available product.

ClearClass doesn't attempt to replace school scheduling systems.

It solves a much narrower problem: making the information those systems already produce useful to students and parents.

Users can upload a PDF, an image or plain text. AI interprets the timetable and transforms it into a structured schedule.

Users can review the result, correct individual details such as lesson names or classrooms, and select from a small set of predefined visual styles.

The output is designed to be readable, printable and shareable.

Both Hebrew and English are supported natively.

There are no user accounts. You arrive, provide a timetable, review the result and use it.

That simplicity is deliberate.

The central AI challenge is interpreting documents that don't follow a standard format.

Different schools organize timetables differently. Inputs may be photographs, PDFs or copied text, with variations in layout, abbreviations, orientation and visual quality.

The system must extract the schedule into a structured representation without losing important information.

Users can correct smaller interpretation errors before printing or sharing.

Privacy was another important consideration.

School timetables may include information about children and teachers. ClearClass therefore provides contextual warnings before sharing, as well as information in its privacy policy.

Shared links use generic previews rather than exposing timetable contents in messaging-app previews.

Regular timetables are automatically deleted after 24 hours. Sharing extends the retention period by three days, and users can also request earlier deletion.

These measures reduce exposure, but they don't make shared links private access-controlled documents. The user still needs to understand what they're choosing to share.

All of this sits outside the original AI transformation concept.

It's also the work that distinguishes a functional demonstration from a public product.

AI coding agents handled the implementation while I directed the product, defined the requirements and evaluated the results.

I collaborated with the agents on architecture, testing and quality assurance.

But the project also exposed a weakness in my development process.

As work progressed, agents kept identifying additional things worth adding to the roadmap.

Many sounded reasonable. Not all were necessary for the MVP.

The result was a gradually expanding scope.

I learned that AI agents can be exceptionally productive at generating work. It's the product leader's responsibility to decide which work should not be done.

ClearClass launched in October 2026.

My objective wasn't to build a large business immediately. It was to release a real product, learn from its usage and see whether the market demonstrated meaningful demand.

I've deliberately kept operating costs very low by sharing infrastructure across my projects and keeping AI inference expenses tied to actual usage.

The first signal I'm watching for is steady, sustained daily traffic, even at a modest level.

The more meaningful milestone is a substantial volume of timetable creations and sharing events, together with how quickly that volume builds once steady traffic is established.

At that point, I'll reassess whether ClearClass deserves more development investment.

Until then, it can continue operating without demanding significant attention.

It has already produced value through reusable infrastructure and as a concrete demonstration of my ability to ship a complete product.