Handwritten-first STEM marking

Reclaim your
evenings and weekends

AEMS app — AI-assisted marking interface with annotated exam feedback

AEMS helps university instructors mark STEM exams, homeworks, and lab reports. Built for handwritten work, and just as comfortable with typed or printed submissions. It reads the paper, follows your rubric, and drafts feedback on the PDF. You review every grade before students see it. Built by a professor at KTH Royal Institute of Technology in Stockholm.

Every semester, the same bottleneck

Grading is the most time-consuming part of teaching, and the hardest to scale.

15–30 min

per paper, multiplied by hundreds of submissions every exam period

Grading fatigue

We drift as we grade, readjusting the baseline as we progress. Consistency suffers

Weeks

By the time students receive feedback, they have already moved on to the next topic

Burnout

Talented academics spending evenings and weekends on marking instead of research or teaching

Connect. Mark. Review. Publish.

If you work on your own, start with the local agent. If you run a course team, connect Canvas. Either way, the marking flow stays in one place and you review the output before it leaves AEMS.

1

Set up

Pair the local agent for Personal, or link Canvas for a team workflow. No manual file juggling once the setup is done.

2

Draft marks

AEMS reads each submission, applies your rubric step by step, and drafts marks with colour-coded annotations. Same criteria, every paper.

3

You Review

See proposed marks and annotations on the PDF. Accept, adjust, or override. Nothing reaches students without your sign-off.

4

Publish

Grades, annotations, and personalised feedback are sent back to Canvas automatically. Students see exactly where they earned or lost marks.

Built for how academics actually work

Not another ed-tech demo. A tool by a professor who has graded thousands of exams.

Academic and teaching assistant reviewing exam submissions together

PDF-native

Works directly with your scanned exam papers. No special apps, no format gymnastics. Upload the PDF, get an annotated PDF back.

Rubric-driven consistency

Your marking scheme, applied identically to every submission. No more drift between the first exam and the last.

AI proposes, you decide

Every mark can be reviewed, adjusted, or overridden before students see it. You are always in charge. AEMS helps with the work, but you remain the examiner.

Native Canvas integration

Submissions flow in from Canvas, grades and comments flow back, all automatic. Students receive annotated feedback directly in their Canvas inbox. No file downloads, no uploads, no copy-paste.

No model shopping to get started

Personal uses the standard AEMS setup with the local agent. Larger deployments can be configured more deeply if your university needs that, without changing the marking flow.

Choose where the PDFs live

Keep source PDFs on your machine with the local agent, use the hosted team workflow, or run AEMS inside your own environment.

Math-aware equation reading

Equations are extracted as LaTeX, not approximate text. A formula-recognition model runs on math-heavy pages, so integrals, fractions, subscripts, and Greek symbols reach the rubric in the form the student actually wrote.

Contextual grading for long-form work

For homeworks and lab reports, AEMS can grade the submission as a coherent document instead of one rubric line at a time. Marks are anchored to specific pages and paragraphs for review.

Student data privacy is not optional

AEMS was designed from day one for institutions that take data protection seriously.

Secure data infrastructure protecting student information
GDPR compliant

EU data residency with documented processing agreements

Read the GDPR report →
No training on your data

Student submissions are never used to train AI models, backed by contractual guarantee

Local-first option

Keep source PDFs on your machine via the paired agent, or deploy on-premises behind your firewall

Full audit trail

Every marking decision is logged, tamper-evident, and traceable for 7 years

Fair & transparent

Rubric-traced marks, override audit trails, and confidence scoring

Our fairness commitment →
Your infrastructure, your rules

On-premises deployment keeps everything behind your firewall

Built for STEM. Works across disciplines.

AEMS is strongest where grading pain is worst: handwritten engineering, math, and physics work (exams, homeworks, and lab reports) with multi-step reasoning and partial credit. It also works for other subjects, but STEM is where it shines.

Individual academic reviewing exam papers at their desk

Individual Academics

“I mark 200 mechanics exams every semester by myself.”

Create an account, pair the local agent, and mark from your own machine. Personal is built for instructors who want faster marking without giving up control of the source PDFs.

See Personal plan →
Stack of exam papers representing large-scale department grading

Departments & Teaching Teams

“We have 6 TAs marking 800 calculus submissions, and consistency is a constant fight.”

Use our hosted platform. Everyone marks against the same rubric, with shared calibration. Push grades to Canvas when the team is satisfied.

Department plan →
Nordic university campus building

Universities & Institutions

“We need this across engineering faculties, behind our firewall, with full audit compliance.”

Deploy AEMS on your own infrastructure. You control the environment, the data, and the access policies. We provide the software and support.

Institutional plan →

Ready to reclaim your evenings?

Your next exam period does not have to mean two weeks of late nights. See which plan fits, or try it on a sample exam today.