- Short lessons work best: one topic, 5 to 10 minutes, one deck.
- Your speaker notes become the brief for the narration, so a few sentences per slide pays off.
- Always review the AI-written script for accuracy before generating audio; you are the subject expert.
- Every export includes captions and a transcript, which helps students who need them and everyone revising.
- When a slide changes next term, edit that slide and regenerate it, instead of re-recording the lecture.
- The same lesson can be narrated in another language for students and families who need it.
The short answer
To turn lecture slides into a narrated video lesson, upload the deck to an AI narration tool, let it write a narration script for each slide, review that script, and export the video with captions. The AI does the writing and the voice. You do what only you can do: check that the explanation is right.
This guide is for teachers and lecturers who want video lessons for a flipped classroom, for absent students, or for revision, without recording themselves for an evening every time a slide changes.
Why turn slides into video lessons at all
- Students can rewatch. A recorded explanation is there the night before the test, not only on the day you gave it.
- Absent students catch up. A lesson video covers the class someone missed.
- Class time goes further. In a flipped classroom, students watch the explanation at home and use class time for practice and questions.
- Captions help everyone. Students who are deaf or hard of hearing need them; students revising in a noisy place use them too.
The usual obstacle is time. Recording narration over 30 slides takes an evening, and fixing one mistake means recording that part again. AI narration changes the cost: a fix is an edit to one line.
Recording it yourself vs AI narration
| Record over your slides | AI narration | |
|---|---|---|
| Time for a 20-slide lesson | An evening, plus retakes | Under an hour, most of it reviewing |
| Fixing a mistake | Re-record that section | Edit the line, regenerate that slide |
| Updating next term | Record again | Change the slides that changed |
| Your own voice | Yes | Yes, with voice cloning, or pick an AI voice |
| Captions | Make them separately | Included with every export |
| Another language | Record it again, if you speak it | Choose a narration language |
Recording yourself still makes sense for a lesson with a personal story, or when students simply need to hear you. For explanation-heavy lessons you will reuse, AI narration saves the most time.
Prepare the deck
- One topic per lesson. Aim for 5 to 10 minutes of video. Split a long lecture into several short lessons; completion drops sharply on long videos.
- Add speaker notes. Two to four sentences per slide on what students should understand. The narration is written from the notes and the slide, so this is where you put your explanation. See how to turn speaker notes into narration.
- Split build slides. Animations do not carry over to the video; each slide becomes a still scene. A slide that reveals a derivation step by step should become one slide per step.
- Check the vocabulary. Note any term with a tricky pronunciation: a scientist's name, a chemical, a place. You can fix it once later.
- Use Google Slides? Download the deck as .pptx, not PDF, so the speaker notes come with it. See turning Google Slides into a narrated video.
Step by step
Step 1: Upload the deck
Start a course or training video project in Unmixr and upload the .pptx (a .ppt or PDF works too). Choose Narrated course video as the output.
Step 2: Tell the AI who the lesson is for
Add a short guidance note: "Year 9 students. Plain language. Define each new term the first time it appears." This shapes how every slide is written, so a lesson for 14-year-olds does not sound like one for graduate students.
Step 3: Choose the language and the voice
Choose the narration language and a voice. A calm, clear voice at a steady pace suits most lessons. To narrate in your own voice, clone it once from a short sample, with your consent.
Step 4: Review the script
Choose Review in Scene Studio. Unmixr writes the narration for every slide, and nothing is charged until you generate the audio. Read every line next to its slide, as you would mark a student's work:
- Is the explanation correct and complete?
- Is it pitched at the right level for this class?
- Are numbers, units and names right?
- Does each slide's narration match the order of what is on the slide?
Step 5: Generate and listen
Generate the audio and watch the first few scenes. Add a pause before an important point, slow down a definition, or add a pronunciation rule for any term that comes out wrong. The rule applies to every slide in the lesson.
Step 6: Export and post
Export the 720p MP4 with its SRT and VTT captions and transcript. Then:
- Google Classroom: attach the MP4 (or add it to Drive and link it) in an assignment or material.
- Canvas, Moodle, Blackboard: upload the MP4 and its caption file to the course page, or a SCORM package on paid plans if your LMS tracks completion that way.
- YouTube: upload as unlisted, add the caption file, and share the link with your class.
Keeping lessons up to date
The real saving shows up next term. When one slide changes, update that slide, edit its narration, regenerate that scene and export again. The rest of the lesson keeps the audio you already reviewed. A recorded lecture would need re-recording; a narrated deck needs one edit.
Lessons in more than one language
The same lesson can be narrated in a different language from the slides, from 100+ languages. That helps students learning in a second language, and families who want to follow what their child is learning. Choose the narration language when you create the project; for whole courses in several languages, see course video localization.
Next step
Take one lesson you teach every year and turn it into a narrated video. A free account exports lessons up to 3 minutes with a small watermark, so you can try it on a real deck before you decide. For a full course, see the AI training video maker.