Today we’re launching Ethyra — an AI grading agent for teachers and teaching teams.
This is v0, our first public release. We built it for a problem that keeps showing up in classrooms and in the research: assessment is consuming the people who deliver it.
The problem
A recurring assessment burden sits on top of an already unsustainable workload. Teachers spend too long evaluating student work. Students get feedback after the learning window has already closed. The assessment pipeline — collect, score, comment, return — is often the worst part of the job, and the least improved part of education software.
That pattern isn’t anecdotal. Recent teacher research, including the Learnosity U.S. teacher survey (2025) and RAND’s State of the American Teacher (2025), keeps pointing to the same pressure: evaluation work expands, instructional time shrinks, and feedback arrives too late to matter as much as it should.
Speed-only tools don’t fix this. Auto-send grades. Bulk-apply comments. Clear the queue. That can look efficient and still fail the classroom — because the hard part isn’t typing a score. It’s judgment: reading work against a rubric, holding a consistent standard, and writing feedback a student can use while the learning is still active.
What we’re building
Ethyra starts from a clear rule: AI should carry the first draft of grading work, not the last word.
You bring the rubric, materials, and grading rules. Ethyra drafts criterion-level feedback across essays, code, handwriting, and mixed submissions. It learns your style as you review. Then you decide what students see.
The loop is simple on purpose:
- Draft — Ethyra proposes scores and comments mapped to your rubric.
- Review — Edit in place. Ask the assistant about a student or a pattern. Adjust tone and standards.
- Release — Nothing reaches students until you confirm.
Alongside that, Ethyra surfaces class and assignment analytics so struggling students and criterion-level gaps show up early — not after the term is already decided.
What’s in v0
- Rubric-based grading for essays, code, handwriting, and mixed work
- Feedback shaped by your teaching style and grading rules
- A reviewing center so teachers keep final control
- Analytics to flag who’s struggling and where
- Classroom privacy defaults: student work isn’t used to train models
v0 is focused. We shipped the grading loop teachers need first — accurate drafts, editable feedback, and a real review step — before adding every adjacent feature.
Why it matters
Better education doesn’t come from removing teachers from assessment. It comes from removing the busywork that keeps them from teaching.
When evaluation takes nights and weekends, standards slip under volume, and feedback lands after students have moved on. When the first pass is drafted in your voice and you stay in the loop, more students can get careful feedback while there’s still time to use it.
That’s the product we’re launching. Not a black box that replaces instructors — a grading agent that works beside them.
Ethyra is live. If you want grading software that assumes the teacher keeps the final say, start with v0.