AI Text

Moodle AIText question type

Release 2.03 Jul 2026

Sample responses are now XML-escaped on export. Any &, < or > in the sample response text previously produced invalid XML that failed to re-import.

Sample responses are deleted when the question they belong to is deleted, rather than being left orphaned in the database.

The AI prompt shown by the "Show prompt" button when previewing a question is now sanitized before output. The renderer previously inserted the raw prompt straight into the page HTML, so script tags, event handlers or javascript: URLs held in the prompt would execute in the previewer's browser. It is now passed through format_text() with cleaning enabled, which strips those while leaving safe markup intact. Thanks to Paola Maneggia for feedback on this release and much other work.

Release 2.02 Apr 2026

Huge thanks to Philipp Memmel and the team at ByCS who have contributed a transformative amount of code and testing to the plugin.

New prompt template system with configurable structured templates using placeholders ({{role}}, {{questiontext}}, {{aiprompt}}, {{markscheme}}, {{response}}, {{language}}). The central template is defined in settings and applied consistently across all AI calls.

Improved MathJax / LaTeX rendering: mathematical expressions in AI feedback are now correctly displayed when the response contains LaTeX notation.

Improved robustness of AI completion against unusual or adversarial student inputs that could previously disrupt prompt formatting.

Fixed the prompt tester in the question editing form so that the loading spinner is removed when an error occurs, preventing the UI from appearing stuck.

Release 1.01 Dec 2025

When editing a question the Sample response evaluation will now render any HTML
rather than showing raw tags.

The "Information for Graders" field towards the end of the question editing form was silently discarding anything added to it. It now saves that information and will appear when manually grading a question.

Thanks to Norbert Forster for reporting that issue in
https://github.com/marcusgreen/moodle-qtype_aitext/issues/51
and thanks to to Thomas Schoenlein for the code to fix this.

Thanks to Philipp Memmel for multiple commits to improve parsing and sanitising of LLM responses and for constant support and feedback with the development of the plugin.

Thanks to Michael Hughes for a fix to the handling of the Tiny MCE Editor.

Thanks to Mathieu Pelletier for a fix which addresses three reported issues
https://github.com/marcusgreen/moodle-qtype_aitext/pull/47
https://github.com/marcusgreen/moodle-qtype_aitext/issues/8
https://github.com/marcusgreen/moodle-qtype_aitext/issues/46

Release 1.00 of the Moodle AIText question type Oct 2025

Confirmed compatibility with Moodle 5.1 by running automated tests
Fixed behat test that was broken by a change of case in a string

Release 0.05 of the Moodle AIText question type Sep 2025

Removed the code that was trimming any extraneious curly braces.
This particularly broke any programming code.
Deal with divide by zero scenario in grade_response.

Feature:

Question text slug

If the slug [[questiontext]] is added to the prompt, the text of the question will be interpolated into the prompt. So if the questiontext is

"Describe an elephant"

The prompt could be

Does this answer the question "[[questiontext]]

Per question response translation

Thanks to Adam Jenkins for this idea (and many others). If a language string is included, e.g. [[language=js]] it will be converted to the string "translate the feedback to the language jp". If the prompt contains [[language=""]] no language translation string will be included. See
https://github.com/marcusgreen/moodle-qtype_aitext/issues/20

Release 0.04 Aug 2025

Merge of callbacks code for upcoming mebis local_ai_connect features.
Fix upgrade.db which was missing the spellcheck fields and breaking the plugin.

Release 0.03 May 2025

Fixed compatibility with Moodle 5.0, thanks to Philipp Memmel of Mebis-lp for the hint about Dependency Injection.

In question editing form changed sample answer allowing multiple sample responses. This saves multiple responses
to be tested against the prompt. Make the spinner prettier for the web service call in the editing form.

Fix backup/restore and xml import/export

Release 0.02 Dec 2024

Merged in code from https://github.com/mebis-lp/moodle-qtype_aitext
Who have have extensive experience with live use of the question type.

It is now possible to select from three different "back end" systems for connecting to the LLM. These are

• Tool AI Connect


• Moodle 4.5 AI Subsystem


• Local AI Manager

Tool AI Connect was the only way previously. The Moodle 4.5 AI subsystem only supports OpenAI ChatGPT and Microsoft Azure (which supports the Open AI Models).

Local AI Manager support Openai (ChatGPT), Ollama, and has other interesting features including this from the readme

https://github.com/mebis-lp/moodle-local_ai_manager/blob/main/README.md

"The AI manager provides the key feature of being a multi-tenant AI backend which is the main reason you should be using this plugin for providing AI functionality in your moodle instance instead of the moodle core AI subsystem."

Removed the ability to pick a different model for each question.

New configuration options in settings.php

Marks prompt in the editing interface is optional

Translation of the response is now optional

Big thanks to Philipp Memmel for help with the switchable backend settings to work and for testing and many other things.

Thanks to Farah Ahmad of Aga Khan University for reporting an issue with plagiarism plugins.

Thanks to Matt Metzgar for reporting an issue with the tranlation postfix.

Release 0.01 May 2024

Version 2024050300

Many thanks to Justin Hunt of Poodll fame https://poodll.com/moodle for contributing code
to allw the testing of prompts from within the question editing form.

Many thanks to Alexander Mikasch of Moodle.NRW (https://moodlenrw.de/) for reporting a prompt injection issue

Refined the default prompt settings to ensure Ollama/mistral returns a number when grading

The model field in the plugins settings can now accept a comma separated list. If there
is more than one model the question editor form will show a dropdown with available models. The selected model is written to the database and will be used as part of the connection to the AI system.

Add model name to disclaimer if configured in settings as [[model]]

Many thanks to Peter Mayer for fixes to coding standards and for code to make the prompt field required. Also for being generally very encouraging.