Training videos that run as long as the lesson does
Write the module, render it in your own face, up to thirty minutes.
Already have an account? Sign in
Free to start · No credit card
Training videos that run as long as the lesson does
Write the module, render it in your own face, and change a line next quarter without booking anyone for anything. One video can run up to thirty minutes on a paid plan, which is the whole module rather than a clip of it.
One render, made from a written script. Nobody booked a room for it.
Kyndrify renders the video files and hands them to you. It does not host your course, run quizzes, or export a SCORM package.
Part of AI video generator.
What is an AI training video generator, and what can it actually make?
A training video generator turns a written module into a finished video of a real person teaching it. You build a Digital Twin once from a photo, a short voice recording and a consent clip, then render each module from its script. One video can run up to thirty minutes on a paid plan. It does not import slide decks or export SCORM.
Who it is for
The person who owns the training library and is also the person on camera. One trainer, or a small team with a Digital Twin each.
What you need
A photo, about ten seconds of clear audio, a consent recording of 15 to 90 seconds, and the module script you have already written.
What it costs
152 credits a second on the paid rate, so a ten-minute module is 91,200 credits. Free starts with 8,400 credits, once, and they do not expire; a free render runs up to one minute.
What it does not do
No SCORM or xAPI package, no quizzes, no completion tracking, no slide or document import, and no screen capture inside the Studio.
Looking for the audience view rather than the tool? Education and training covers who this is for and what a library looks like without a studio.
How long one video can run, and what it costs
The ceiling on a paid plan is thirty minutes for a single video. At about 150 words a minute that is roughly a 4,500 word script, and you hand it over in one piece: it is read in sentence-aligned chunks and comes back as one file. On Free, each render runs up to one minute.
Here is the part most tools will not print. Video is billed by the second, so a module costs arithmetic rather than a quote:
- one minute, 9,120 credits.
- three minutes, 27,360 credits.
- five minutes, 45,600 credits.
- ten minutes, 91,200 credits.
- thirty minutes, 273,600 credits.
Set that against what a plan gives you each month. Plus is 28,000 credits, about 3.1 minutes of finished video. Pro is 56,000 credits, about 6.1 minutes. Max is 123,000 credits, about 13.5 minutes. Ultra is 307,000 credits, about 33.7 minutes.
So the ceiling is real, and almost nobody should spend a course that way. One thirty minutes render is 273,600 credits, which is more than a whole month on Max. Cut the course into modules of three to six minutes, which is where learners stay anyway, and the same budget covers far more of the curriculum. The ceiling is there for the module that genuinely needs it, not as the way to work.
Two more numbers worth knowing. Every video with a person on screen bills a minimum of 120 credits, which no module of this length comes near. And renders are capped per day as well as by credits: three a day on Free, ten on Plus and fifty on Pro, with the plans above them higher again.
Three steps, and the first one happens once.
- 1
Build your Digital Twin.
Add a photo and about ten seconds of clear audio, then record a consent clip of 15 to 90 seconds. The face in it is checked against your photo, and that recording is the only filming you ever do.
- 2
Write the module and render it.
Paste the script, pick your voice, and the finished video comes back with every word spoken. Long scripts are fine: a thirty minutes module goes in as one script on a paid plan, and the cost is on the screen before you press anything.
- 3
Hand it to your LMS.
Download the video, run it through the Transcript tool for an SRT or a WebVTT file, and upload both wherever your course already lives.
What comes back.
One person, the whole module.
A finished module is a file: one person talking, the words you wrote, and the same face and voice as the last one you made. Nothing has to be assembled afterwards, and nothing about the picture drifts between module three and module eleven. Paid renders come back at 1080p; a free render is 720p and carries a small visible mark, which comes off from the Plus plan and above.
The week the policy changes
The first recording is not the expensive part of a training library. The sixth revision is.
What usually happens
- The module was filmed in March, in a room you no longer have.
- In September one sentence about approvals stops being true.
- Re-recording it needs the same room, the same light, the same shirt and the same person.
- So a note goes in the course description instead, and the video keeps saying the old thing to every new starter.
What happens here
- You open the script and change the sentence.
- You render the module again, with the same face and the same voice, because they are your Digital Twin and not a booking.
- The old file stays in your Workspace until you replace it.
- You pay for the module's seconds, not for a day.
Five kinds of training video
The same tool, five different script shapes. These are the ones training teams actually ship.
New-hire onboarding.
Day one, before anyone has met anyone. Ninety seconds each: who you are, what this team is for, what to do first, and who to ask when you are stuck. Make one per team rather than one long induction, because the ops module and the sales module go stale on different clocks.
Tutorials and SOP walkthroughs.
The how-to for the process people keep getting wrong. Keep the spoken part to the decisions, and leave the clicking to a screen recording. There is no screen capture inside the Studio, so record the software with the tool you already use, bring that file into the editor, and top and tail it with your Digital Twin. The person explains, the screen shows.
Compliance refreshers.
The annual one nobody looks forward to. Write it once, render it once, and re-render the paragraph that moves when the rule moves. Read the note further down before you write this one.
Product and process updates.
The short module that says what changed and what people do differently from Monday. These are the ones that justify the whole setup, because they are only worth making if making them is cheap.
Announcements.
The reorganisation, the policy, the new system. One person, two minutes, saying the thing plainly instead of a paragraph in an all-staff email that half the company reads as bad news.
Captions and a transcript for your LMS
A finished module gives you three files if you want them.
The video downloads as an mp4, and it is yours to upload wherever the course lives.
The transcript comes from the Transcript tool: run the module through it and download plain text, SRT or WebVTT. Cues are cut the way subtitles should be, at most 42 characters a line, two lines a cue, and no cue longer than five seconds. Transcribing costs 1 credit a second of audio with a floor of 20 credits. Most learning management systems take an SRT or a VTT file beside the video, and that is the sidecar they want.
Burned-in captions are a different thing, and they live in the editor, which is open on every plan. If the module will be watched on a phone with the sound off, burn them in; the export hands back an SRT and a VTT file as well. If it lives in a course player that draws its own captions, use the sidecar and keep the picture clean.
What we do not make is the package. There is no SCORM or xAPI export, no quiz, no completion tracking and no hosting. Your learning management system does that part, and it does it better than a video tool would.
One module, more than one language
Take a module you rendered here, pick a supported language, and it comes back re-voiced in your own voice. Speech and dubbing work in ten languages. It costs 80 credits a second of the original with a floor of 120 credits, and it is on the Plus plan and above.
The scope is worth being exact about, because it is the question everyone asks second. This works on a video that was made in Kyndrify, not on a file you upload, so there is no door for dubbing the module you recorded somewhere else. One run does one language, which means three markets is three runs from the same original.
A note on compliance training
Compliance modules are a big share of what training teams make, so here is the honest boundary.
Kyndrify renders the video. It does not tell you what your training has to cover, it does not certify anything, and nothing here satisfies a regulation on its own. Whether a module meets an obligation is a question for whoever owns that obligation where you work, and it is the same question it was when you filmed the module on a camera.
Two things are ours to say. A Digital Twin renders only against a recorded consent, and the face in that recording is compared with the face in the photo, so a named person is never used in a module they did not agree to. And built-in AI labelling is still rolling out across output types, so if a learner should be told how the module was made, write that line yourself rather than assuming the file carries it.
What this does not do
No SCORM or xAPI package.
No quizzes, no completion tracking, no course hosting. You get video files and caption files, and your learning management system does the rest.
No slide, PDF or document import.
There is no door that turns a deck into a course. You bring the script, or you plan one with Kae, who prices the plan first.
No screen capture inside the Studio.
Record the screen with the tool you already use, then bring that file into the editor and cut it together with the rendered module.
Not a live class.
These are recorded modules. Nothing here runs a session with people in it, and nothing posts or schedules them for you.
Questions training teams ask first
- How long can one training video be?
- Up to thirty minutes for a single video on a paid plan. At about 150 words a minute that is roughly a 4,500 word script, and it goes in as one piece. On Free, each render runs up to one minute, which is enough to see what your Digital Twin looks like before you commit to anything.
- What does a ten-minute module actually cost?
- 91,200 credits, at the paid rate of 152 credits a second. For context, Plus gives you 28,000 credits a month and Max gives you 123,000. That is why most training libraries are built as modules of three to six minutes rather than one long video, and it is the honest answer rather than the flattering one.
- Can I import my slide deck, or export a SCORM package?
- No to both. Kyndrify imports no slide deck, document, weblink or course package, and exports none of them either. You write or paste the script here and render the video; your learning management system handles packaging, tracking and hosting.
- What happens when the content changes?
- You edit the script and render the module again. Your Digital Twin is the same face and the same voice as the first version, so a corrected module does not look like a patch, and you pay for the module's seconds rather than for a day.
- Can I get captions or a transcript for the LMS?
- Yes. Run the finished module through the Transcript tool and download plain text, SRT or WebVTT, then upload the sidecar beside the video. Transcribing costs 1 credit a second of audio with a floor of 20 credits. If you would rather burn the captions into the picture, do that in the editor instead.
- Can more than one person present?
- Yes. Each trainer builds their own Digital Twin, with their own consent on record. Plus holds 3 characters of your own, Pro 10, Max 25 and Ultra 50. Team seats are on the Pro plan and above at $20 a person a month, with five included on Ultra, so a training team works inside one Workspace.
The tools that do this
More on this from the blog
Write the module. Render the module.
Build your Digital Twin from a photo and a short recording, then render your first module from a script you have already written. Free to start, with 8,400 credits that do not expire.