Workflow Slides, PDF or script → narrated MP4

AI Training Video Maker for slides and scripts

Unmixr turns a slide deck or a training script into a narrated MP4 with AI narration, captions and a transcript, ready for your LMS. Upload a PowerPoint, PDF or Word file, pick a voice, and export: there is nothing to record and nothing to install.

Free: export up to 3 minutes with a watermark; paid plans remove both. Plans from $9/month.

Demo: a PowerPoint fire-safety deck becomes a narrated training video in Unmixr Real product demo · 2:37

Fire safety is everyone’s job.

Unmixr makes a training video in four steps: upload your slides or script, choose an AI narrator, review the narration, and export a 720p MP4 with SRT/VTT captions and a transcript. A 10-slide deck is usually ready in a couple of minutes, in a web browser, with no microphone and no video editor.

How to make a training video with AI in four steps

  1. 1. Upload your slides or scriptUpload a .pptx, .ppt or .pdf deck, or a Word document with your script. Each slide or page becomes a scene, and its speaker notes (or the slide text when a slide has none) become the draft narration.
  2. 2. Choose a narratorPick an AI voice by language, tone and teaching style, or an approved cloned voice. The slide language is detected for you, and different scenes can use different voices for role plays.
  3. 3. Review the narrationEdit any line, set how acronyms and product names are pronounced, add pauses, and retime scenes in Scene Studio before anything is rendered.
  4. 4. Export and publishDownload a 720p MP4 with SRT/VTT captions and a transcript, or a SCORM package on paid plans, then dub the same module into other languages when you need to.

Free accounts export up to 3 minutes with a short Unmixr watermark; paid plans remove both. Decks from Google Slides or Keynote work too: export them as .pptx first.

Real product demo · 2:37

Watch a slide deck become a narrated training video.

One real fire-safety deck, start to finish: upload the slides, let AI write the narration from the speaker notes, change the voice in Scene Studio, and export the MP4.

Make your training video free
  • Free credits to start
  • Your speaker notes become the script
  • MP4, captions or SCORM
eLearning & training

Upload the deck. Publish the course.

A real clinical training deck, dropped into Unmixr and published as a narrated video lesson — no recording booth, no editing timeline, no re-record when the slides change.

Input PowerPoint deck 4 of 36 slides
Title slide of the Caesarean Scar Ectopic training deck
Definition slide from the training deck Causes slide from the training deck Ultrasound imaging slide from the training deck
Caesarean_Scar_Ectopic_Training.pptx 1.7 MB · the untouched source file
Download the deck
Output Narrated training video MP4 · 720p
Frame from the narrated Caesarean Scar Ectopic training video showing an ultrasound imaging slide
17:53 AI narration
Narrated end to end with an Unmixr AI voice Every line stayed editable before export

Start from slides PPTXPPTPDF

Or from a document PDFDOCXEPUBTXTMarkdownHTML

1 Your source becomes the script

Upload a deck or a document — Unmixr reads it and drafts narration section by section. Rewrite any line before a word is voiced.

2 Pick the teaching voice

A voice for every use case, with pitch, speed, and emphasis you can tune per sentence or per term.

3 Captions and transcript

SRT and VTT subtitles plus a full transcript come out of the same project — ready for your LMS.

4 Localize the course

Dub the finished lesson into 100+ languages without rebuilding a single slide.

What you get with every training video

One project produces the video, the accessibility files, the LMS package, and the other-language versions. Nothing has to be stitched together in a second tool.

A narrated MP4

A 720p (1280x720) video with one scene per slide, each timed to its narration. It plays in any LMS, on YouTube or Vimeo, in Teams or Slack, or straight from a shared drive, and it is free of watermarks on every paid plan.

SRT and VTT captions

Caption files generated from the same script the narrator reads, so the words on screen match the words spoken. Upload them beside the video for learners on mute, in a noisy office, or who simply prefer to read along.

A full transcript

The complete narration as text, for reviewers who sign off on wording, auditors who need to see exactly what a compliance module said, and learners who search a lesson instead of scrubbing through it.

A SCORM package

On paid plans, a course project can be packaged as SCORM 1.2 or SCORM 2004 for any SCORM-compliant LMS, with completion tracked from what the learner actually watched rather than how far they dragged the playhead.

Dubbed language versions

The same module narrated or dubbed in 100+ languages, with translated subtitles. Translation is a text step you review before anything is voiced, so product names and terminology arrive intact in every version.

Narration audio and an editable project

The voice track on its own for podcast-style lessons or another editing tool, plus the project itself: change one line next quarter, regenerate that scene, and export again without touching the rest.

What you can turn into a training video

Start from whatever the training already exists as. Sources with pages become a narrated video; a plain script becomes narrated audio with captions.

.pptx / .ppt PowerPointNarrated video One scene per slide, split automatically. Speaker notes become the narration; slides without notes are narrated from their text.
.pdf PDF deck or handoutNarrated video One scene per page. A PDF carries no speaker notes, so the draft narration is written from what is on each page.
Google Slides Google Slides exportNarrated video In Google Slides choose File, Download, Microsoft PowerPoint (.pptx), then upload the file. Speaker notes come across with it.
Keynote / Canva Keynote or Canva exportNarrated video Export the deck as PowerPoint (.pptx) and upload it. In Keynote that is File, Export To, PowerPoint.
.docx Script in a Word documentNarrated video A written script or SOP becomes video scenes built from the document's structure, each with editable narration.
Paste / .txt / .md / .html Plain scriptNarrated audio A pasted or plain-text script has nothing to show on screen, so it becomes narrated lesson audio with captions and a transcript.
URL Web page or help articleNarrated video Paste a link into the Blog or Article to Video workflow and the page's text becomes a narrated video with captions.

Source files can be up to 1 GB and 1,500 pages. Not sure which you have? Upload it: the wizard reads the file first and only offers video output when the source has pages it can render.

Who uses the AI training video maker

Built for L&D teams, enablement managers, HR, and educators: anyone whose training already lives in slide decks and documents, and who has to keep it current.

HR and onboarding teams

Turn the onboarding deck every new hire sits through into a narrated module they can watch on day one, at their own pace, in their own language. When the handbook changes, the module changes the same afternoon.

E-learning voiceover →

Compliance and safety trainers

Annual compliance, fire safety, data protection and code-of-conduct modules are slide decks that change every year. Edit the lines that changed, regenerate those scenes, and re-export with a transcript auditors can read.

PowerPoint to video →

Product enablement and customer education

Ship narrated feature walkthroughs for sales, support and customers as fast as the product changes. Pronunciation rules keep product names right in every video, and the same module can go out in every market you sell in.

Course narration →

Course creators and online schools

Turn lecture slides into narrated lessons with one consistent instructor voice across the whole course, then sell it in more languages without re-recording. Every paid plan includes commercial use.

Course narration →

Healthcare and clinical educators

Clinical teaching decks are dense with terms a generic reader gets wrong. Set the pronunciation of each drug, procedure and acronym once; the 17-minute clinical lesson on this page was published from an unedited 36-slide deck.

See the clinical sample →

Global and multilingual teams

Roll the same module out to every region: narrate it again in the target language with the same instructor voice, or dub the finished video with translated subtitles, from one project.

Course video localization →

AI narration vs recording a training video yourself

Recording yourself and hiring a voice actor both work. The difference shows up the second time, when a slide changes or a region needs its own version.

Question Record it yourself Hire a voice actor Unmixr
First draft of a 10-slide module A quiet room, a microphone, several takes, then audio editing and syncing to slides. A brief, a booking, and a turnaround measured in days. A couple of minutes after the upload, playable in the browser.
Changing one slide later Re-record the line and try to match the original tone and room sound. A new booking, or a pickup session at the talent's rate. Edit the line and regenerate that scene; the rest stays as approved.
Same voice across 20 modules Depends on the same person being available and sounding the same every time. Possible while the same actor stays available. One narrator for the whole curriculum, with shared pronunciation rules.
Other languages Find a speaker per language and repeat the whole process. A separate actor and session per language. Narrate again or dub in 100+ languages from the same project.
Captions and transcript A separate captioning pass. Usually a separate captioning pass. SRT, VTT and a transcript with every export.
What you need Microphone, recording space, audio and video editing software. Budget, scheduling, and review cycles. A web browser and the deck you already have.

How to write a training deck that narrates well

The narration is only as good as the script behind it. These six habits make the first draft close to the final one.

Write the narration in the speaker notes

Speaker notes become the script, so a deck with good notes produces a good first draft. Write them the way you would say them out loud: short sentences, one idea each, and the word you want stressed near the end of the sentence.

Keep slides visual, put the words in the voice

Learners cannot read a paragraph and listen to a different one at the same time. Leave the slide to a headline, a diagram or three short bullets, and let the narration carry the explanation.

Aim for 30 to 60 seconds a slide

That is roughly 75 to 150 spoken words. A slide that needs three minutes of narration is usually three slides, and splitting it gives learners a visual change at the moment the idea changes.

Split animated builds into separate slides

Animations and transitions are not carried into the video, because each slide is rendered as a still image. If a build reveals four steps, make it four slides; the narration then lands on each step as it appears.

Fix pronunciation before the first render

List the acronyms, product names, drug names and people's names in the deck and add a pronunciation rule for each. A rule applies everywhere the word appears, so it is a one-time fix for the whole module.

Have a subject-matter expert read the transcript

Reviewing text is faster than reviewing video. Send the transcript for sign-off, apply the corrections in Scene Studio, and regenerate only the scenes that changed before you publish.

What to know before you start

The plain facts, so the first video matches what you expected.

  • Narration comes from speaker notes first and slide text second; you can rewrite every line before rendering.
  • PowerPoint animations and transitions are not preserved. Each slide becomes a still scene timed to its narration.
  • Videos export as 720p MP4. SRT/VTT captions and a transcript come with every export.
  • Free accounts export up to 3 minutes with a short Unmixr watermark; paid plans remove both.
  • Starter narrates up to 80 slides per project; Pro and Business take longer decks, from files up to 1 GB.
  • SCORM packaging is available on paid plans. Commercial use is included on every paid plan.
  • Nothing to install: Unmixr runs in the browser on Windows, Mac, Linux and Chromebooks.

Free to start, plans from $9/month

Make and hear a real module on a free account. Upgrade when you want full-length, watermark-free exports.

Lite $9/month 100K credits every month
Starter $19/month 500K credits every month
Pro $49/month 2.6M credits every month
Business $99/month 6.5M credits every month

Every paid plan includes commercial use, watermark-free exports and SCORM packaging; annual billing gives you two months free. The free account exports up to 3 minutes with a short watermark, and the 7-day trial starts in one click.

FAQ

Training video maker: your questions, answered

Straight answers for teams evaluating AI training videos.

What is the easiest way to make a training video with AI?

The easiest way to make a training video with AI is to upload the slide deck you already have to Unmixr, pick an AI narrator, and export a narrated MP4. Each slide becomes a scene, the speaker notes become the narration, and the captions and transcript are generated with the video. A 10-slide deck is usually a finished draft in a couple of minutes, with nothing to record and nothing to install.

What is Unmixr?

Unmixr turns slides and scripts into narrated training videos, with AI voices in 100+ languages, subtitles, transcripts, dubbing and SCORM output, from $9 a month. The training video maker is the workflow that does it end to end: slides or a script go in, and a narrated, captioned MP4 (or a SCORM package) comes out of one browser-based project.

Can I turn a PowerPoint into a training video?

Unmixr turns a PowerPoint deck (.pptx or .ppt) into a narrated training video, with one scene per slide. A slide's speaker notes become its narration; when a slide has no notes, the slide text is used instead. Every line stays editable in Scene Studio before you render, so you can rewrite the script without touching the PowerPoint file.

Can I make a training video from a script instead of slides?

A script becomes a narrated training video when you upload it as a Word document (.docx), which Unmixr splits into scenes from the document's structure. A script pasted into the wizard, or uploaded as .txt, .md, or .html, has nothing to show on screen, so it becomes narrated lesson audio with captions and a transcript rather than a video. If you want visuals, put the script in the speaker notes of a simple deck and upload that.

Does it work with Google Slides, Keynote, or Canva decks?

Google Slides, Keynote, and Canva decks all work in Unmixr once they are exported as a PowerPoint (.pptx) file. In Google Slides use File, Download, Microsoft PowerPoint; in Keynote use File, Export To, PowerPoint. A PDF export works too, but a PDF carries no speaker notes, so the narration is drafted from the slide text.

Do I need to record my own voice or own a microphone?

You do not need to record anything or own a microphone, because every line of narration in an Unmixr training video is generated by an AI voice from your text. You choose a narrator by language, tone, and teaching style and audition it on your own script. If a real instructor's voice matters, you can clone an approved voice with the speaker's consent in 80+ languages.

What files do I get when I export a training video?

Every Unmixr training video export includes a 720p MP4, SRT and VTT caption files, and a text transcript. The narration can also be downloaded on its own as audio, course projects can be packaged as SCORM on paid plans, and the same project can be dubbed into other languages with translated subtitles.

Does Unmixr export SCORM packages for an LMS?

Unmixr packages course and training projects as SCORM 1.2 or SCORM 2004 on paid plans, so a module drops into any SCORM-compliant LMS as a trackable course. Completion is tracked from the parts a learner actually watched rather than a furthest-position pointer, so skipping to the end does not mark a module complete. The plain MP4 with captions remains the option for platforms that prefer video files.

Are PowerPoint animations and transitions preserved?

PowerPoint animations and transitions are not preserved, because Unmixr renders each slide as a still image timed to its narration. If a slide relies on a build, split it into two or three slides that each show one step; the narration then moves with the visual. Scene timing, background music, and the voice for each scene can all be adjusted in Scene Studio.

What resolution are the training videos?

Unmixr exports training videos as 720p (1280x720) MP4 files. That size keeps slide text sharp on a laptop screen and streams smoothly in an LMS player or on a slower office connection, and the MP4 plays in standard video players and on YouTube, Vimeo, or an intranet.

Can I make a training video for free?

A free Unmixr account makes real training videos: exports run up to 3 minutes and carry a short Unmixr watermark, and paid plans remove both. The free account comes with 5,000 credits, which is enough to narrate a short module and hear it before you decide anything. The narration plays in the studio before you export, so nothing is spent on a video you have not heard.

Is there a free trial?

The Unmixr free trial lasts 7 days and includes 5,000 credits (+20,000 more when you add a card); nothing is charged during the trial, and you can cancel anytime before it ends. Start it in one click from the app. Adding a card during the trial also unlocks clean, full-length exports, so you can publish a real module before the first invoice.

How much does it cost to make training videos with Unmixr?

Unmixr plans start at $9 a month for Lite with 100K credits a month; Starter is $19, Pro $49, and Business $99 a month, and every paid plan includes commercial use and watermark-free exports. One credit is one character of narration, so a ten-minute module of about 1,500 words uses roughly 9,000 credits. Annual billing gives you two months free.

How long can a training video or slide deck be?

One Unmixr project narrates up to 80 slides on Starter and longer decks on Pro and Business, from source files up to 1 GB and 1,500 pages. A typical slide narrates for 30 to 60 seconds, so an 80-slide deck is roughly an hour of training video. On longer decks a slide picker lets you choose which pages become scenes, and long curricula work best as one project per module sharing the same narrator and pronunciation rules.

How long does it take to make a training video?

A typical 10-slide training deck becomes a narrated video draft in about one to two minutes, and longer decks scale roughly with slide count. Rendering runs in the background, so you can keep working. Most of the real time goes into reviewing the script, pacing, and pronunciation, which is the part worth a subject-matter expert's attention.

How do I update a training video when the content changes?

To update an Unmixr training video, edit the lines that changed and regenerate only those scenes; the rest of the video keeps the narration your reviewers already approved. Re-export and the captions and transcript update with it. A policy change or a renamed menu item becomes a five-minute edit instead of a new recording session.

Can every module keep the same narrator?

One AI voice can narrate every module in an Unmixr curriculum, so lesson twelve sounds like lesson one. Pronunciation rules and pacing settings carry across the project, and a module you add next year can use the same narrator. Where you have permission, an approved trainer's voice can be cloned so the course keeps a real instructor's identity.

How do I get acronyms, product names, and technical terms pronounced correctly?

A pronunciation rule in Unmixr fixes how a term, acronym, or product name is spoken once, and the rule applies everywhere that word appears in the project. You can also slow a passage down, add a pause around a definition, and stress the words that carry the learning objective, so a dense clinical or technical module stays clear.

Can I translate a training video into other languages?

Unmixr can narrate the same training module in another language or dub a finished training video, with subtitles, in 100+ languages. Translation is a text step you review before anything is voiced, so names and terminology survive. The Course Video Localization page compares the two paths, narrate again or dub, side by side.

Can I use the training videos commercially or for client work?

Every training video you make on a paid Unmixr plan is cleared for commercial use, including internal training, paid courses, and client deliverables. The one rule that always applies is consent: only clone a voice you own or have written permission to use.

Do Unmixr training videos help with accessibility requirements?

Unmixr covers the media side of accessible training: every video exports with SRT or VTT captions and a full transcript, and pacing controls keep the narration clear. Standards such as WCAG apply to the whole course, so treat the captions and transcript as components of that work rather than a certification of it.

Can my team work on the same training library?

Business plans give a team shared credits and seats, so instructional designers, subject-matter reviewers, and voice editors work from one account and one library. For larger seat counts, get in touch from the pricing page.

Turn your next training deck into a video today

Upload the slides you already have, pick a narrator, and hear the first module before you decide anything else.

Make a Training Video Free

Free account to start · Plans from $9/month · 100+ languages

Still have a question?

If you have a specific use case or need help configuring a custom workflow, our team is always here for you. We would love to listen to your needs!