Gobinar turns your texts into a complete course: chapters and lessons, slides with images, word-for-word narration, voice-over and a finished video for every lesson. You watch, and correct wherever you want to.
Example structure. Chapters, lessons and videos come out of your own material.
Courses rarely fail on content. They fail on everything between "I know this" and "it exists as a video": outline, slides, recording, re-takes, editing, uploading, structuring it in a learning system. Recording thirty lessons one by one is not an afternoon, it is a project — and it is the reason so many sold courses keep their buyers waiting.
The same goes for onboarding inside a company. The knowledge sits in people's heads and in the documentation. What is missing is someone who turns it into a structured course in reasonable time — and updates it again at the next release.
Gobinar takes exactly that part off your hands. You bring the knowledge, we build the lessons.
Texts, links, a product — or a finished script. Gobinar can also read web pages when all you have is the address.
The AI works its way into your topic and gathers what belongs in the course.
Chapters, lessons and learning objectives arrive as a proposal. You reorder, rename, throw out.
Slide design, colours, voice, presenter image — once for the whole course.
For every lesson: slides, narration, voice-over, finished video. Runs in the background while you do something else.
The course sits ready at Gobinar — with a video per lesson, playable and replaceable at any time.
Not a script for you to record yourself, but a lesson you can play. For 200 credits each.
Title, content and a fitting image per slide — in the course design.
What is said on each slide, word for word and editable.
Voice-over in a voice of your choice, your own if you want it.
Cut and finished at 1920×1080 — ready to play, ready to export.
If you have already written the narration, it does not get "improved". It is cut into slides along your stage directions and a word budget — by a fixed rule, not by a language model. The AI then only adds what is visible.
For everyone who has to control exactly what is said — finance, health, law, accredited training.
Your portrait photo as a still image in a circle costs nothing. If you want one, you get a presenter whose lips move in time with the text. Anything generated by AI is marked inside the video file itself — exactly as the European AI Act requires from August 2026.
Just give us the link to a sales or info page — Gobinar works out what it is about. Internal addresses are blocked; this is not an invitation into other people's networks.
An existing LearnDash course can be pulled in: structure and texts come along, the videos are then produced here.
Every finished lesson exists as its own video and can be played straight away. You are not tied to any one player.
You do not pay for a course, you pay for the steps that actually run. Before each one, its price is shown.
A course with 24 lessons therefore lands at roughly 5,100 credits.
Video services bill by render minute. At the volume a course produces, that quickly becomes the largest item on the whole bill.
Measured on our production server, 25 slides, four CPU cores. Not an estimate.
For the same amount of render time. That is why our rendering runs on our own hardware — and why a lesson costs 20 credits instead of many times that.
Summary, slides, checklist and a QR code back into the course — generated automatically from what is there anyway.
Questions drawn from the lesson content, each with its source, for you to approve before any participant sees it.
Course, chapters and lessons including video move to WordPress with LearnDash at the press of a button. The route is built and runs — but today it hangs on one fixed installation and cannot yet be set up for your own system.
We list this because it is coming — and explicitly as "not yet", because nobody can buy it today. A subscription taken out on a promise is not a subscription.
If the course is meant to explain how your own software works, it must invent nothing. That is why this route does not start from text someone types, but from your repository and your documentation — and every statement in the course points back to its source.
Example recording. The figure in it is from the first real run, 14 August 2026.
The GitHub connection requests read permissions only — not a single write permission. Nothing is cloned, built or executed.
Before anything is taken apart, the material is scanned for keys and passwords. A single hit blocks the entire run — and the report never names the value itself.
Every piece of evidence is produced by a fixed rule, not by a model: file, line range, verbatim quote, checksum. Only then does the AI turn it into lessons.
What gets read: README files, Markdown files and interface descriptions. Code, test and dependency directories are skipped — your source code is neither read nor processed.
Whatever the AI claims is checked against the evidence that actually exists. Anything not pointing at evidence is discarded — not corrected. If a lesson ends up without evidence, the draft cannot be approved.
The draft goes into an approval view. Video production starts only after your yes.
First real run on 14 August 2026 against a real repository: 25 documentation files, 10,103 pieces of evidence, 37 seconds.
This route is not clicked, it is discussed — it depends on your repository, your permissions and your data protection rules. We enable it per company, together with you.
Arrange a conversationNo account needed, no mailing list — we answer within one working day.
There is no separate course plan you would have to buy on top. What you build spends credit — and how much credit arrives each month is set by your plan.
Only if you want it to. If you supply a finished script, it stays word for word — it is merely cut into slides along your stage directions and a word budget. The AI then supplies only what is visible: heading, bullet points, image idea. That matters for anyone who has to control exactly what is said, for legal or professional reasons.
It runs in the background, lesson by lesson, and you can watch it happen. The rendering itself is fast: a twelve-minute video is produced on our server in under three minutes. The AI steps before it take longer, because that is where the actual thinking happens.
Yes, and precisely. Swap out individual slides, regenerate an image, redo a voice-over, change the narration — without rebuilding the whole course.
Not at the press of a button yet. The handover to WordPress with LearnDash is built and works — but it currently hangs on one fixed installation and cannot yet be set up for your own system. What does work today: every finished lesson exists as its own video and can be played, so you can embed it in any platform that accepts video. We would rather say so plainly than let you find out while trying it.
200 credits per lesson, plus a one-off 180 for the research and 120 for the structure. A course with 24 lessons therefore lands at roughly 5,100 credits. Before every step you see what it costs.
Create a course, drop in your material, look at the structure. That costs nothing — and a single finished lesson tells you whether the result meets your standard.
No credit card · cancel any time · servers in Europe