Heymans AI tutor

Copyright 2024-2026 Sebastiaan Mathôt (@smathot), Wouter Kruijne (@wkruijne), University of Groningen

Heymans AI tutor is a Python library and web app for LLM-based teaching tools. The initial focus will be on automated grading of open exams and interactive quizzes.

Dependencies

For Python dependencies, see pyproject.toml. Python 3.10 or later is required. In addition to these, a local redis server needs to run for persistent data between sessions. pandoc needs to be installed for generating invidual feedback reports.

Using as Python library

See example/heymans-brightspace-example.py for an example of how to grade an open-ended exam. This example assumes that you're using the Brightspace learning environment. The example folder also contains example output that is generated during grading.

Running as web server (development)

Download the source code, and in the folder of the source code execute the following:

pip install .               # install dependencies
python app.py               # start the app

Next, access the app (by default) through:

http://127.0.0.1:5000/app/quiz

License

Heymans is distributed under the terms of the GNU General Public License 3. The full license should be included in the file COPYING, or can be obtained from:

• http://www.gnu.org/licenses/gpl.txt