AI-powered marking for real-world assignments.

MarkMate reads your PDFs – assignments, tests, treatises and more – and generates structured, rubric-aligned feedback in minutes instead of hours.

Launch MarkMate
10x
Faster marking
Consistent
Rubric-based feedback
Multi‑doc
Batch processing support

Built for educators who work with PDFs every day.

MarkMate is a focused tool for marking real assignments – not a generic chatbot. It understands rubrics, scores consistently, and fits into your existing workflow.

Upload once, mark in batches

Upload individual PDFs or ZIP archives of a whole class. Organize submissions into batches and keep track of their status from upload to final report.

Rubric‑driven scoring

Use existing rubrics or generate new ones with the AI rubric generator. MarkMate aligns feedback and scores directly with each rubric criterion.

Rich feedback, not just a grade

Students get structured comments, strengths and weaknesses per criterion, and optional language feedback on clarity, grammar, and referencing.

PDF annotation & reports

Choose between annotated PDFs with comments in context, or standalone assessment reports summarizing scores and feedback for each submission.

Support for exams & MCQs

Handle essays, long‑form assignments, treatises and even MCQ answer sheets. Re‑run marking with different strictness levels when you need to adjust.

Training & transparency

Export anonymised marking data to refine your own models or audit AI decisions. MarkMate is designed to help, not replace, human markers.

How MarkMate fits into your workflow.

From PDF upload to final report, in four simple steps.

  1. 1. Upload your PDFs

    Drag‑and‑drop assignments, tests, treatises or ZIP archives. MarkMate extracts text automatically, even from scanned PDFs via OCR.

  2. 2. Choose or generate a rubric

    Select an existing rubric, or use the built‑in generator to create criteria and point allocations from sample answers or syllabus documents.

  3. 3. Let the AI mark

    Pick your AI provider (OpenAI or Anthropic), strictness level, and output type (annotated PDF or separate report). MarkMate runs criteria‑by‑criteria marking.

  4. 4. Review, adjust, and export

    Use the results dashboard to review marks, re‑run specific submissions, export CSV summaries, or download annotated PDFs and reports for sharing.

Designed for busy educators and programme coordinators.

Whether you are marking a stack of undergraduate essays, postgraduate treatises, or continuous assessment tasks, MarkMate helps you save time while keeping your academic standards.

  • Lecturers and tutors marking large cohorts
  • Programme coordinators needing consistency across markers
  • Assessment units managing exams and written scripts
  • Educators exploring AI to augment, not replace, human judgement

Why not just use a chatbot?

Generic chat tools are not built for high‑stakes assessment. MarkMate:

  • Stores your data in your own database
  • Uses explicit rubrics and scoring rules
  • Tracks versions of marking results
  • Can be tuned per course and institution

Frequently asked questions.

Is MarkMate a hosted service or self‑hosted?

MarkMate is designed to be self‑hosted. You run it on your own server (for example, using Virtualmin as documented), connect it to your own database, and control how data is stored and accessed.

Which AI models does MarkMate use?

Out of the box, MarkMate supports OpenAI and Anthropic (Claude) via their official APIs. You configure API keys in the environment and can switch provider per marking run.

Can I still mark manually?

Yes. There is a dedicated manual marking interface where you can capture scores and feedback by hand, or adjust AI‑generated marks before finalising them.

How do I access the app?

Once deployed, the MarkMate app runs at /tools on your domain. For example: https://markmate.io/tools. Use the “Launch MarkMate” button above to go straight there.

Ready to see MarkMate in action?

Log in to your instance and upload a real assignment. You can be looking at AI‑assisted feedback in under 10 minutes.