From the pile to handing back, in five steps you can follow
Laura assists you at every step of the correction. You make the call on the grade: transparent, auditable and with no silent cloud upload.
The correction flow
- 01
Upload the pile
Photograph or scan the papers. The images stay on your device at first.
Details for the technically curious
Technical: local-first by default. No image leaves your device before you deliberately release it. - 02
Laura reads the handwriting
Laura reads locally and for free first. Only when something stays unclear does she also ask a stronger AI, targeted at the uncertain spots.
Details for the technically curious
Technical: 4-stage cascade. Two local tracks form a ROVER consensus (free); on uncertainty it escalates to cloud AI with Writer-Adaptive Recognition. Hard cases go straight into the more precise track. - 03
Laura makes a proposal
Following your assessment criteria, Laura proposes points and comments, always with evidence. It stays a proposal, not a grade.
Details for the technically curious
Technical: assessment along your rubric, with evidence, a confidence value and alternatives. - 04
You review and decide
You confirm, change or discard every proposal. Anything uncertain, Laura puts in front of you visibly: she does not guess, she asks.
Details for the technically curious
Technical: review arena with a queue. Spots with low confidence are gathered and made visible. - 05
Done: correction on the paper plus PDF
Your correction appears right on the paper, in your own handwriting. You export as PDF only once every single assessment is truly final.
Details for the technically curious
Technical: inline overlay in the authorised teacher font; export is released only when every graded decision is final (No-Surprises-Gate).
What Laura does fundamentally differently
The teacher decides
Laura never sets a final grade on her own. Every assessment is a Teacher Final Decision.
Free and local first, then AI when needed
The cascade saves cost and data: cloud only when the local consensus is not enough, and then pseudonymised (opt-in).
Laura learns the handwriting
The more Laura reads from a pupil, the more reliably she recognises their handwriting (technical: Writer-Adaptive Recognition).
Individualisation: the pedagogical best case
The real promise is not speed, but coverage. Because every paper is seen with the same thoroughness, an evidenced picture emerges for every child. That way, no one simply slips through the system any more.
Every paper equally thorough
The AI does not tire. The twentieth paper gets the same attention as the first. Assessment drift across the pile is made visible and checked, not ignored.
Error patterns instead of single mistakes
Laura clusters what a child works on repeatedly. Many red marks turn into one concrete support hint.
Strengths become visible
Because every spot is mapped to the rubric, the teacher sees where a child is already secure. That is the basis for nurturing special talents instead of losing them in the average.
Time for those who need it
The time regained flows to the weakest and the strongest. Exactly where teachers are missing and classes are large, the effect is strongest.
Ready to win back correction time?
Start on your own or bring your whole staff. We will set up an account for you.