An AI-powered system that analyzes Norwegian medical documentation and generates exam accommodation recommendations for university students.
Students with documented disabilities are entitled to exam accommodations at Norwegian universities, but manual case review is slow and inconsistent. This tool helps coordinators by automatically parsing medical records, identifying relevant diagnoses and functional limitations, and mapping them to standard accommodation categories — all in seconds.
Supported universities: UiO, UiB, NTNU
The student selects their university so recommendations are tailored to that institution's specific accommodation rules and deadlines.
Drag and drop (or select) Norwegian medical documents. Supported formats: PDF, JPG, PNG, HEIC (max 10 MB).
The system extracts diagnoses and functional impact from the documents and maps them to concrete accommodation categories — such as extended time, technical aids (Lingdys, word processor, electronic dictionary), separate exam room, or reading assistance.
Each recommendation is shown with a confidence score. The coordinator (or student) reviews the list and confirms which apply before the report is finalized.
The system also scores the uploaded document for authenticity based on metadata, signatures, and identity signals — flagging missing HPR-IDs, absent author fields, or OCR-processed files that may obscure metadata.
- Runtime: Node.js
- Language processing: Norwegian NLP
- Deployment: Docker / Docker Compose
- API: OpenAPI 3.0
# Clone the repo
git clone https://github.com/jakobkoding2/tilrettelegging-bot.git
cd tilrettelegging-bot
# Start with Docker Compose
docker-compose upSee DEPLOYMENT_GUIDE.md for full setup instructions.
Norwegian universities are legally required to offer exam accommodations to students with documented disabilities. Manual assessment of applications is time-consuming and prone to inconsistency between reviewers. This system assists coordinators by providing evidence-based AI recommendations derived directly from the submitted medical documentation, making the process faster and more uniform.