Two ways to use AEMS
A hosted grading portal for teaching teams, and an open-source local agent for academics who want local file storage with a hosted account workflow.
AEMS Agent
Open SourceFor individual academics · Personal plan
An open-source local bridge used by the Personal plan. It keeps your source exam PDFs on your own machine and exposes them to your authenticated AEMS account when you choose to grade.
How it works
- ✓ Runs on Windows, macOS, and Linux
- ✓ Pair it with the hosted AEMS app for Personal-plan local storage
- ✓ Connect Ollama locally or via Ollama Cloud
- ✓ Source exam PDFs stay on your machine via the paired agent
Why open source?
The agent is open source so you can verify exactly what it does. It stores files locally and serves them to your authenticated AEMS session on localhost. Account authentication and access enforcement happen in AEMS services, not in the agent. Inspect the code yourself.
AEMS App
All plansThe grading portal · Every plan includes it
The AEMS App is the same web-based grading portal on every plan. Build rubrics, grade exams, generate annotated feedback, and export results. What changes between plans is where your data is stored and whether you can collaborate with others.
Same app, different storage
- ✓ Personal — files stay on your machine via the Agent
- ✓ Department — EU-hosted file storage, encrypted at rest, 30-day auto-purge
- ✓ Institutional — self-hosted on your own infrastructure, behind your firewall
Team collaboration (Department & above)
Higher-tier plans add an organisation with shared file hosting. Multiple teaching assistants work on the same course files, mark against the same rubric, and calibrate together — ensuring consistent grading across the whole team.
- ✓ Shared rubrics and cross-marker calibration
- ✓ Canvas LMS integration: pull submissions, push grades
- ✓ Roles and permissions per course
Full control (Institutional)
Deploy the entire AEMS platform on your premises. Connect your own local LLM engine so no data ever leaves your network — not even to an external AI provider.
- ✓ Connect a local LLM — all AI inference stays on-site
- ✓ Full audit logs for every grading action
- ✓ Custom branding with your institution’s identity
- ✓ User admin panel — manage accounts, roles and access centrally
- ✓ Multi-tenant architecture — isolate faculties, schools or departments
- ✓ Grading queue — batch submissions and process at scale
- ✓ Your firewall, your data retention policies, your rules
At a glance
| AEMS Agent | AEMS App | |
|---|---|---|
| What it is | Open-source local agent that bridges your files to the hosted app | Web-based grading portal (same on every plan) |
| Source code | Open source (GitHub) | — |
| File storage | Your machine only | Your machine (Personal), EU cloud (Dept), or your server (Institutional) |
| AI provider | Ollama (local or cloud) | Personal: Ollama. Department: managed by AEMS. Institutional: institution-controlled. |
| Team collaboration | Single user | Organisation with shared files, rubrics, calibration (Dept & above) |
| Audit logs | — | Full grading audit trail (Institutional) |
| Custom branding | — | Your institution’s identity (Institutional) |
| User admin | — | Central account & role management (Institutional) |
| Multi-tenant | — | Isolated faculties / departments (Institutional) |
| Grading queue | — | Batch processing at scale (Institutional) |
| Canvas LMS | Manual export | Direct integration |
| Offline use | Yes (with local model) | Yes (Institutional with local LLM) |
Not sure which to pick?
If you grade alone and privacy is paramount, start with the Agent. If you coordinate a team of markers, use the App. Both produce the same annotated feedback.