Concept buildA worked example of how we would build this for a typical business. Not client work: no client, no client data, no achieved results.
AI Video
Concept buildA concept build for a pipeline that turns a teacher’s lesson outline into a captioned explainer video, with the teacher’s review at the centre of every step. A model drafts the script, the teacher edits and approves it, and slides, captions and the render follow automatically before the video is uploaded to the LMS.
The brief
A typical edtech content team in India has more lessons to explain than it can film. Each explainer passes through scripting, slide design, recording, editing and captioning, and the subject teacher waits at every handover. Production capacity, not teaching quality, decides how many lessons get a video, and captions are often added last or skipped. Courses taught in Hindi and English double the work. The team wants teachers to stay in charge of what is taught while the repetitive steps between an approved script and a finished, captioned video stop needing a person.
Explainer
Explainer · 7 slides · about 44 s
Concept buildConcept build: Lesson Explainer Video Pipeline for an EdTech Team
Concept build, explained in 7 slides: a concept build for a pipeline that turns a teacher’s lesson outline into a captioned explainer video, with the teacher’s review at the centre of every step.
Slide 1 of 7
01 · AI video generation
Lesson Explainer Video Pipeline for an EdTech Team
Concept build: Lesson Explainer Video Pipeline for an EdTech Team.
02 · The problem
What a typical EdTech content team is up against
A typical edtech content team in India has more lessons to explain than it can film.
03 · The approach
How we would build it, in 5 steps
We would build it in 5 steps, from lesson format and slide templates to lMS upload and versioning.
04 · Architecture
How the pieces connect
How the parts connect, from the first input to the person who reviews the result.
05 · Demo · sample run
One lesson, from outline to the LMS
A sample run, step by step, including where a person steps in.
06 · Typical targets
What we would measure it against
Typical targets, each with the basis it would be measured against.
07 · Next step
Scope a build like this for your business
Book a strategy call to scope a build like this.
Architecture
Architecture of the Lesson Explainer Video Pipeline for an EdTech Team concept: Lesson outline → Script model (outline); Script model → Teacher review (script draft); Teacher review → Slide templates (approved script); Teacher review → Caption builder (approved script); Licensed media → Slide templates (images); Slide templates → Remotion render (scenes); Caption builder → Remotion render (caption track); Remotion render → LMS via its API (versioned upload).
Diagram with 8 parts: Lesson outline (source), Licensed media (source), Slide templates (processing), Caption builder (processing), Remotion render (processing), Script model (model), Teacher review (human), LMS via its API (system of record).
Connections:
Sample run
Teacher submits the outlineRoutine
Sample lesson: placing fractions on a number line, class 6, in English and Hindi.
Script draftedRoutine
Narration and slide text in the lesson format; two added examples highlighted.
Teacher review gateA person takes over
The teacher rewrites one example and approves; the approval is stored.
Slides and captions builtRoutine
Templates filled from the approved script; captions generated from the same text.
Render check fails onceNeeds care
One Hindi line overflows its slide; the template shrinks the text and re-renders.
Uploaded to the LMSRoutine
Video, caption track and version 1 attached to the lesson through the API.
Build plan
We agree a lesson format with the academic lead: learning objective, worked example, common mistake, recap and practice prompt. That format becomes a small set of Remotion slide templates in the platform’s style, with Devanagari and Latin fonts embedded so Hindi and English text renders correctly at any resolution.
The teacher writes or pastes a lesson outline, and a language model drafts the narration and on-screen text in the agreed format, at the course’s reading level. The draft highlights every factual statement it added beyond the outline, so the teacher can see exactly what the model introduced.
Nothing renders until the teacher approves the script on a simple review screen, editing lines directly or asking for a redraft with a note. The approved script, not the model’s draft, is the source for every later step, and the approval is stored with the teacher’s name and the date as the record of who signed off the content.
The approved script fills the slide templates, and captions are generated from the script itself rather than from speech recognition, so they match the words exactly. Narration is either the teacher’s own recording against a timing guide or, if the team prefers, a synthetic voice labelled as synthetic in the video. Remotion renders each lesson in the course’s formats, under a Remotion company licence where the team’s size requires one.
Finished videos upload to the LMS through its API, with captions attached as a separate track, the lesson ID and a version number. When a teacher updates a script, the pipeline re-renders and replaces the video in place, keeping the earlier version in storage so the change history stays with the lesson.
Typical targets
Typical target
A reviewable first draft within one working day of the outline
Basis: Draft, layout, captions and render are automated with teacher review kept; timed per lesson in the pilot
Typical target
Production time per lesson cut from days to hours
Basis: The team’s current production steps are timed at scoping, then compared lesson by lesson in the pilot
Typical target
Captions on every video, matching the approved script word for word
Basis: Design property: captions are generated from the script, not transcribed from the audio
Typical target
No video published without the teacher’s approval
Basis: Design property: rendering starts only from an approved script, and the approval is stored with the lesson
Your project
Footage, voices, music and fonts checked for licence and use before they enter a template.
Your type, colours, intro and outro built once as reusable templates, so every video stays on brand.
Scripts are signed off before render and the final cut before publishing; nothing goes out automatically.
Aspect ratios, caption files and file sizes matched to the channels you actually use.
Where this fits
FAQ
No. The pipeline is a concept build, and the fractions lesson in the demo is a sample written to show each step. For a real edtech team in India the starting point would be timing how long a lesson takes to produce today, because the time targets on this page only mean something against that baseline.
The teacher. The model only drafts, anything it adds beyond the outline is highlighted, and nothing renders until the teacher approves the script. That approval is stored with the lesson, so a named person has always signed off the content, and corrections go through the same review.
No. The pipeline handles lesson outlines, scripts and media only, never student records or learning data. That keeps children’s personal data, which India’s DPDP Act protects with parental-consent rules, out of the system entirely, and the LMS service account only needs permission to upload course content.
Yes, and for many teams that is the better choice. The teacher records narration against a timing guide generated from the approved script, and the render lines the slides up with the recording. A synthetic voice is available for teams that want one, labelled as synthetic on screen; cloning a real teacher’s voice is outside this concept’s scope.
The teacher edits the approved script and approves the change, and the pipeline re-renders only that lesson and replaces it in the LMS under a new version number. The previous version stays in storage with its approval record, so the team can always show what was taught and when.
Worked examples
Concept builds are worked examples, not client work.
Concept buildAI Video
A concept build for a pipeline that turns a listing’s own data and licensed photos into short vertical and widescreen videos on one brand template. A model writes scene captions only from the listing fields, every number on screen is checked against the record, and nothing is published until a reviewer at the portal approves it.
Next step
Book a free 30-minute strategy call. Bring one process you want to automate and we will talk through scope, integrations and a sensible first step.
Free 30-minute call · no commitment
