Autumn-term pilot · UK universities

Marking feedback, at the speed you can say it.

Your markers talk through a piece of work the way they’d talk to a colleague. Markaloud transcribes it and drafts structured, consistent feedback against your rubric, with suggested marks, for the marker to check. Nothing reaches a student until they send it.

See the pilot terms Download the one-page PDF →

Free for the autumn marking period · up to 20 markers · no purchase order, no procurement

How it works

1

Mark by talking

Markers speak while reading — a dissertation, an essay, a lab report, a reflective assignment, a final-year project. No templates, no second typing pass.

2

AI drafts, the marker decides

Speech is transcribed, and the AI drafts feedback against your assessment criteria (strengths, weaknesses, actions) with a suggested mark for each when you use a rubric. The marker can change any of it.

3

Marker checks, student receives

The marker reviews and edits the draft, then sends it. Students get fuller, more consistent feedback than time usually allows.

Any discipline

Rubric-driven, not subject-specific.

Markaloud ships with templates for essays, reports, dissertations, lab reports and presentations — or import your own from Excel, Word or Canvas. Departments in this round span engineering, business, nursing, psychology and education.

The autumn pilot

What

Full Markaloud access for your department for the autumn marking period

Cost

Free. No purchase order, no procurement. If it isn’t working, stop using it.

Markers

Up to 20, onboarded by us in a 30-minute session

Term

September – January marking

You commit to

A named contact, and 30 minutes of honest feedback at term end

After the pilot

If it earned its place: institutional plans from £1,950/year (20 markers). If not, your data is exported and deleted on request.

“You mark aloud. AI drafts. You check it and send it.”

Why departments run it

  • Marking time falls where it hurts most — long-form written work and project reports.
  • Feedback becomes consistent across markers and legible to external examiners.
  • A defensible AI position for your teaching committee: the AI drafts from the marker’s own words and suggests marks, and the marker checks everything before a student sees it.
  • Dictation removes a real barrier for markers with dyslexia, RSI or a visual impairment.
  • Works alongside Canvas; no change to your marks process.

Governance, answered up front

UK GDPR
We sign a Data Processing Agreement with your institution before any student work is uploaded.
AI training
Never. No student work is used to train any model, contractually, by any provider we use.
Where data sits
Student work is stored in the UK. Transcription and structuring are carried out by named sub-processors in the US under a UK transfer agreement. The full list is published and named in the DPA.
Marker control
Markaloud transcribes the lecturer’s spoken marking and drafts written feedback from it, with suggested marks against the rubric when one is used. The lecturer reviews and edits everything before it reaches a student. The overall mark is worked out from the criterion marks by a fixed weighted calculation, not by the AI.
Audit & exit
Transcriptions, feedback runs, account exports and deletions are logged. Opening feedback or playing a recording isn’t logged. Export and deletion on request, at any time.

Worth twenty minutes?

I’ll show you Markaloud working on one of your own rubrics. No slides.

Email me Download the PDF →