After a year of building Assessain, we want to be completely transparent with you about where things stand — and ask for your help deciding what comes next.

Where We Are Today

Assessain has been live for over a year now. We've built a working platform — PDF worksheet creation, AI-powered grading, student submission links, and more. Some of you have been using it regularly, and we're grateful for that.

But here's the honest truth: we haven't found the right product-market fit yet. While the tool works, the feedback we keep hearing is the same:

The AI grading is nice, but it doesn't quite match how I would grade it myself.

And that's a real problem. If you have to re-check and correct every AI grade anyway, the time savings are marginal. We understand that — and we think we know why.

The Core Issue

Every teacher grades based on years of experience and instinct — criteria that go beyond what any written rubric can capture. Things like "this answer is technically correct but misses the point" or "this student is improving, so I give the benefit of the doubt."

Right now, Assessain's AI grades based only on the rubric and a sample good answer. It doesn't know your standards, your expectations, or your grading style. It grades like a generic AI — and frankly, ChatGPT can do that for free.

That's not good enough. We need to offer something meaningfully better.

A Possible New Direction: AI That Learns Your Style

We're considering a major shift in how Assessain works. Here's the idea:

1. You review and correct the AI's grades on the first 3–5 student papers (you'd do this anyway)
2. Assessain learns from your corrections — what you scored higher, what you scored lower, and why
3. The AI then re-grades the remaining papers in your style, not in a generic way
4. Over time, Assessain builds a grading profile for you — the more you use it, the better it gets

The result: instead of "AI grading" you get "AI grading that matches how YOU grade."

This would be something ChatGPT can't do (no memory between sessions), and something no other grading tool currently offers.

The Three Paths Forward

We see three possible futures for Assessain:

Path A: Build the calibration feature. If enough of you believe this would genuinely change how useful Assessain is, we'll invest the next few weeks building it. The technical work is feasible — we just need to know it solves a real problem for you.

Path B: Continue improving the current approach. Maybe calibration isn't the issue. Maybe there are other features or fixes that would make Assessain worth paying for. Tell us what they are.

Path C: Wind down the project. If the honest answer is that AI grading tools just aren't useful enough yet, we'd rather know now than keep building something nobody needs. We'd open-source the code and move on.

We Need to Hear From You

This isn't a rhetorical question — your response will directly determine what happens next. We're asking every Assessain user to take 2 minutes and reply to this email (or write to juda.kaleta@gmail.com) with your answers:

1. Does the "AI that learns your grading style" idea sound valuable to you? Would it change how much you use Assessain?
2. What does the AI get wrong most often when grading your students' work? Can you give a specific example?
3. Would you pay €5/month for an AI grading tool that reliably matched your grading standards?

Even a one-line reply helps. Even "no, I'm not interested anymore" helps. Silence is the hardest thing to work with.

Thank you for being part of this journey. Whatever comes next, your input shapes it.

— Juda, creator of Assessain