Audio in, feedback out
Learners upload an audio file and get a transcript, written AI feedback with strengths and areas to improve, and a score if the activity is graded.
AI Coach — screenshot pending
ai-functions/images/edzaicoach-index-1.pngGiving every learner detailed, consistent feedback on a spoken performance — a pitch, a language exercise, a presentation — takes hours of listening and doesn’t scale.
AI Coach is a Moodle activity where learners submit an audio recording and receive AI-powered feedback. The teacher defines the evaluation criteria; the audio is transcribed and analysed against them, and the learner gets structured feedback and a score that flows into the standard gradebook.
Audio in, feedback out
Learners upload an audio file and get a transcript, written AI feedback with strengths and areas to improve, and a score if the activity is graded.
Teacher-defined criteria
You write free-text evaluation criteria that guide the AI directly, so feedback targets exactly what the task is meant to assess.
Grades in the gradebook
The AI score (0–100) maps proportionally to your grade scale and is written to Moodle — and you can manually override any grade.
You control disclosure
Show feedback immediately, or set it to never so you review first; set maximum attempts and the upload size limit per activity.
Teachers
Add an AI Coach activity, write the evaluation criteria, choose the maximum file size, attempts, when feedback is shown, and a grade. Then open the activity’s submissions list to review transcripts, AI feedback, and scores, overriding grades where needed. See For teachers.
Learners
Open the activity, consent to AI processing the first time, upload your audio, and select Submit for AI Feedback. The page updates automatically when your transcript, feedback, and score are ready — and you can submit again if attempts remain. See For learners.
Administrators
Configure the OpenAI API key, default model, and default file size in Admin settings, and make sure Moodle cron is running.
Setting up (teachers)
Each activity carries free-text evaluation criteria, a maximum file size, a maximum number of attempts, and a feedback-visibility choice (immediately or never). Allowed audio formats are mp3, wav, ogg, m4a, and webm. If the activity is graded, the AI’s 0–100 score maps proportionally to your scale; you can manually override the grade and mark a submission reviewed. With feedback set to never, the grade is still recorded while the feedback text stays hidden until you share it.
The learner flow
First-time use requires consent to send audio to an AI service. After submitting, the submission shows as being analysed and the page updates automatically when ready — usually within a minute or two. Feedback includes a collapsible transcript, written comments, and a score where graded.
Processing and providers
Analysis runs in the background via Moodle’s scheduled-task (cron) system, so cron must be running for feedback to be produced. An administrator sets the OpenAI API key (used for transcription and feedback), the default AI model, and the default maximum upload size.
mp3, wav, ogg, m4a, and webm. The accepted formats and size limit are shown by the upload box in the activity.
The teacher writes free-text evaluation criteria on the activity, and those are used directly to guide the feedback — so being specific helps.
Analysis runs on Moodle cron in the background. If submissions stay in the “being analysed” state, check that cron is running (at least every few minutes) on your server.
Yes. The AI score maps to your grade scale, but you can manually override the grade from the submission view and mark it reviewed.
Yes. The first time they open the activity they’re asked to consent to their audio being sent to an AI service; consent is required to submit.