AI Video Lectures

Scholarly can turn your source material into fully narrated, animated video lectures. Upload PDFs, text files, images, or just describe a topic -- and the platform generates a video with animated scenes, illustrations, charts, and voice narration, complete with chapters, a transcript, and AI chat. A compliance policy becomes a training module, a product spec becomes an onboarding walkthrough, a certification syllabus becomes a lesson you can watch on a commute.

Creating a Video Lecture

Video lecture creation is a two-step process.

Step 1 -- Pick your content. Choose your source material:

• PDFs -- Pick PDFs from your library or upload new files. How many sources one lecture can combine depends on your plan: one on Free, up to ten on Premium, up to twenty on Laureate — the same shared limit every creation tool uses. See Plans and Limits. For long PDFs, use the page selector to pick exactly which pages the lecture should cover. The AI uses extracted text to locate relevant pages, then reads those pages visually when needed so diagrams, equations, tables, and handwriting can ground the generated scenes.
• Text files -- Upload .txt, .md, or .csv files.
• Images -- Upload photos, diagrams, or screenshots.
• Google Drive -- Pick files straight from your Drive, no downloading and re-uploading. See Connections.
• Link -- Paste a link to a website or an online PDF.
• Prompt -- Use the Prompt tab to describe any topic and generate a video from scratch, no files needed.

Step 2 -- Customize settings. Before generating, you can configure:

• Length -- The lecture is built to the runtime you choose, in every mode. Free lectures with a length control are fixed at 2 minutes and the control stays locked. A completed free lecture's persistent Make this video lecture longer? banner can upgrade and remake it at about 10 minutes in the same video without another creation credit; the original stays playable until the replacement finishes. Paid plans can set a target up to 45 minutes before generating, at every resolution, and the lecture actually fills the length you pick rather than compressing into a short recap. Short mode uses its own short video format and has no length control. See Plans and Limits.
• Video Mode -- Choose how the lecture teaches: Standard, Illustrated, Short, Manim Math Animation, or Freestyle. Every mode is available on every plan. On a paid plan a new lecture starts on Freestyle, where the AI reads your source and picks the format that suits it; on Free it starts on Standard. Either way every mode is listed, and the one you pick yourself always wins. See the full Video Modes guide for what each mode does and when to use it.
• Voice -- Open the voice picker to choose your narrator. The list is searchable and every voice has a preview so you can hear a short demo before you commit, or leave it on Auto and Scholarly picks a voice that suits the material. The choice is made per lecture, in this step.
• Video quality -- Choose the resolution your lecture renders at: 720p (the default — crisp and a good balance) or 1080p Full HD (the sharpest, available on paid plans). See Plans and Limits for which qualities your plan can generate.
• AI model -- Auto uses GPT 6 Luna on every plan, at medium reasoning on Free and extra-high reasoning on a paid plan, so a lecture left on Auto costs one AI Creation Credit. Paid users can choose another model when they want a different style, while free users can browse those alternatives with lock badges. Claude Sonnet 5 and Kimi K3 cost 2 AI Creation Credits, while GPT 6 Astra and Claude Opus 5.5 cost 3. See Choosing an AI Model for the full defaults table.
• Customization Profile -- Apply a saved Customization Profile from Custom Instructions to reuse guidance, colors, logos, typography, and visual references across lectures. A selected profile applies to the default Standard lecture as well as the other modes. Profiles work in every creation tool and are available on every plan.
• Language -- Choose the language for narration. Dozens of languages are supported, and the selector uses the same saved account preference as every other AI creation tool. Choose Auto when you want the lecture to match the language of your source material.
• Scholarly logo -- Branding is included by default. Paid plans can turn it off before generating. Watermark-free means generating with the logo turned off and then downloading that file -- downloading cannot strip a logo from a video that was already rendered with one.
• More options -- Three extras sit behind More options in the Appearance section: Dark Mode, which renders every slide on a dark theme; Look, a color theme picker (leave it on Auto to draw a fresh one); and Background music, an optional soft music bed under the narration, off by default. Dark Mode and Look apply to Standard lectures, so they appear when Standard is selected.
• Companion slide deck -- Standard lectures always include one. There is no toggle to turn it on or off.
• Custom instructions -- Add free-text instructions to steer the lecture: emphasize certain chapters, explain for a visual learner, focus on key formulas or common mistakes, and more.

Once you confirm, generation begins. You can close the modal and come back later -- you will receive a notification when the video is ready.

Getting the Length You Picked

A 15-minute pick arrives as a 15-minute lecture, not a 6-minute recap.

Narration is recorded chapter by chapter. The real speaking pace is measured from the first chapter's finished audio, and the chapters still to come are sized to that measurement — so the runtime is steered by what the narrator actually sounds like, not by an estimate made before a word was spoken.

If a plan comes back shorter than your target, it is extended with new scenes drawn from your sources rather than stretched or padded. Nothing is trimmed to hit a number, and no lecture is refused for being long or short.

The one hard stop is a 45-minute ceiling on total narration. A lecture whose script runs past it renders every scene that was voiced, so you get the lecture up to that point instead of losing the whole thing.

What You Asked For, Next to What You Got

Under the player, the finished video page prints the length you requested beside the runtime actually delivered, the pages of the source the lecture was built from when there is a page range, and the language it was narrated in. A lecture that came in shorter than your target is visible as such rather than quietly presented as the whole thing.

A free preview says so too: it is labelled as a preview and shows its own runtime alongside the target length of the full lecture, so you can see exactly what upgrading would produce before you decide.

Generation Progress

Every lecture opens a live progress page the moment you start it. For video it counts the work you actually care about: how many scenes are planned, how many have been recorded, and whether it is currently recording narration or rendering the finished video. A long generation shows real movement rather than one step that appears stuck.

A Standard lecture starts recording its opening chapter as soon as the scene plan is ready, instead of waiting for the whole plan to be reviewed first — the review runs alongside that recording and folds into the scenes not yet recorded. Longer picks are planned chapter by chapter in parallel. A 12-minute lecture now reaches its first narration in roughly half the time it used to.

Short lectures often take 10-25 minutes. Long targets can take more than an hour, especially at 1080p, because narration and every video frame still have to be generated and validated. You can cancel an in-progress generation from its progress page, or retry a failed one. If an older lecture was built on a model that is no longer offered, retrying re-runs it on your plan's current model instead of asking you to start a new request.

If a Lecture Can't Be Created

A failed lecture says what went wrong -- the content was declined, a source couldn't be read, a source was too large, the sources didn't hold enough to work from, or the script, narration, or rendering step broke -- and offers only the fixes that apply to that reason: Try again, Try another model, Pick fewer pages, Replace the source, or Adjust the sources.

When the cause lives in the source itself, a bare Try again is deliberately not offered, because replaying the same sources would fail the same way. Your sources and settings are restored into the panel, so you are not rebuilding the request from scratch, and a creation that fails returns its AI Creation Credit automatically. See When a Creation Fails for every reason and fix.

Video Quality

You choose the resolution your lecture renders at in the customize step. Higher quality looks sharper but takes a little longer to generate and produces a larger file.

Quality only affects how sharp the picture is — every video, at every resolution, still comes with the same chapters, captions, interactive questions, transcript, and AI chat. Which qualities you can generate depends on your plan; see Plans and Limits for the breakdown.

Scenes and Visuals

Each video is composed of multiple scenes. The AI picks the scene type that best fits what's being explained:

• Definitions -- A term on screen, written like a textbook entry, with the explanation building up alongside it.
• Comparisons -- Two ideas side by side with their differences highlighted.
• Step-by-step processes -- Numbered stages that animate in as the narrator walks through each one.
• Equations -- Math rendered with proper typography so symbols, fractions, integrals, and superscripts read cleanly.
• Stats and figures -- Animated count-ups for big numbers and key statistics.
• Charts -- Bar charts, line graphs, and pie charts drawn live as the narrator describes the data.
• Key-insight callouts -- Short, emphasized takeaways for the most important moments in the lecture.
• AI-generated illustrations -- Custom imagery for concepts that benefit from a visual.

The visual style itself is designed to feel like a beautifully laid-out textbook in motion -- warm paper background, serif headlines, clean diagrams, and gentle animation. Narration is paired with natural-sounding AI text-to-speech in the language and voice you picked.

How Much a Standard Lecture Shows

A long Standard lecture is built from many short scenes rather than a handful of long ones — a 45-minute lecture plans around three scenes a minute — so the picture changes with the explanation instead of holding one card for five minutes.

Scenes that do run long reveal their content in stages, timed to the narration, so something on screen changes every 10 to 15 seconds.

On-screen text is sized to how much a scene has to show, equations are laid out to fit within the frame rather than running past its edges, long labels wrap onto a second line instead of shrinking to an unreadable size, and light-mode cards read with more contrast. A card that used to recur as a refrain through a whole lecture now appears once per chapter. A packed policy document or a technical spec stays legible all the way through.

Chapters and Transcript

Videos include chapter markers so you can jump to specific sections. A full transcript of the narration is available on every completed video, with search support so you can quickly find specific moments or topics.

Watching a Lecture

The player has the controls you would expect and a few worth knowing about:

• Picture-in-picture — pop the video out into a floating window and keep it playing while you work elsewhere.
• J and L — jump ten seconds back or forward.
• Playback speed — set it once and it follows your account, so your laptop and your phone agree.
• Chapters — click any chapter marker to jump straight to that section.
• Captions — toggled from the control bar; see below.

A lecture reopens at the second you stopped, on any device you sign in from, with a Start over if you would rather begin again. See Continue Where You Left Off.

If playback stays stuck buffering for about half a minute, a Retry appears so you can reload from your place instead of waiting it out. Brief buffering, a deliberate pause, and time spent in a background tab do not trigger it.

Companion Slide Deck

Standard lectures come with a companion slide deck, generated alongside the video. Open it in the Slides tab on the video page, or use the three-dot menu and choose Download slide deck to save a copy.

Captions

Every completed video lecture ships with synced captions, on every plan including Free. The captions button sits in the video player's control bar — click it to show captions, click it again to hide them. They follow the narration line by line, so you can read along while you watch, follow a lecture with the sound off in an open-plan office, or scan a long lecture at 1.5× or 2×.

Captions are written from the same script that drives the narration, not transcribed from the finished audio, so they match word for word what the speaker says and land on the right frame. They are generated in the language you created the lecture in — a lecture narrated in Japanese comes with Japanese captions out of the box, with no separate setup. Regenerate the lecture and the captions regenerate with it.

If you download the MP4, the captions stay available in Scholarly's player as a separate caption track rather than being burned into the picture.

Interactive Questions

Every video lecture comes with built-in checkpoints. At key moments a small, non-intrusive question card slides up at the bottom of the video — playback keeps rolling, you hover or tap to expand it, answer, and the AI gives you instant feedback. You can turn the cards off anytime with the Interactive questions toggle in the panel below the player.

Each full pass through the questions is saved, so you can see your Best and Last scores at a glance and open your full attempt history right next to the questions list. You can also create flashcards from the whole video without leaving its page. See the Interactive Video Questions guide for the full breakdown.

Background Music

Background music is an optional toggle under More options in the customize step, off by default. Turn it on for a subtle music bed under the narration — a faint, atmospheric layer rather than something competing with the speaker. The mix keeps the voice front and center, even at lower playback speeds.

Creating from an Upload

You can also create a video lecture when you upload files. Add your files, open the More menu, and choose Create Video Lecture. A customization panel lets you pick a length, a video mode, an AI model, the language, and custom instructions before generating.

Creating from AI Chat

You can create video lectures directly from the AI chat. Say something like "create a video lecture on our refund policy" or "turn this into a video lecture" and the AI generates it for you. When the video is ready, it plays inline in the conversation.

AI Chat

Every completed video lecture has a dedicated AI chat panel. Ask questions about the video content, request clarification on a topic, or explore related ideas. The AI uses the video transcript and your original source materials to give grounded answers.

Flashcards

Generate flashcards directly from completed video lectures. Open the flashcard creator from the video page to turn key concepts into cards you can drill.

Sharing

Share any video lecture with a link. Recipients can watch the video, read the transcript, and browse chapters without needing to create an account.

Downloading the Video

On a paid plan, open any completed lecture, click the three-dot menu in the top right, and choose Download video. Scholarly downloads the rendered MP4 exactly as it was generated, including its narration and any background music you enabled. For a watermark-free file, turn the Scholarly logo off before you generate -- downloading cannot remove a logo after the fact. Synced captions remain available in Scholarly's player as a separate caption track. On every plan, you can share a public viewing link for free.

Research Depth

Before writing the script, the AI actively researches your source material -- rereading key sections, pulling out real examples, and finding the most citation-worthy passages. A longer target length -- or a "go deeper" custom instruction -- gives the lecture room for more examples and connections where the material supports them.

Limits

Free users get 1 lifetime AI Creation Credit, shared across video lectures, podcasts, AI Slides, flashcard decks, and other AI tools. It does not reset. Premium includes 10 AI Creation Credits each week; Laureate includes 80 each week. Most models cost 1 AI Creation Credit, Claude Sonnet 5 and Kimi K3 cost 2, and Frontier models cost 3. Every Video Mode is available on every plan; free lectures with a length control are fixed at 2 minutes, and their persistent Make Longer banner can replace the same video at about 10 minutes after an upgrade without another creation credit. Choosing a different target before generation remains a paid feature. One lecture can combine 1 source on Free, up to 10 on Premium, or 20 on Laureate. Paid plans also unlock 1080p Full HD rendering, custom video length up to 45 minutes at every resolution, watermark-free output, and finished MP4 downloads. See Plans and Limits for the full breakdown.

Related

• Video Modes — what each of the five teaching styles does, and which one fits your material.
• Interactive Video Questions — the checkpoint cards built into every lecture.
• When a Creation Fails — every failure reason and the fix that goes with it.
• AI Podcasts — the same sources as a listenable discussion instead of a video.
• Plans and Limits — credits, source limits, lengths, and resolutions by plan.