Guide PDF to Video Lecture

Turn a PDF Into a Narrated Video Lecture

To turn a PDF into a narrated video lecture in Scholarly, open New → Video lecture, upload the PDF (or Word file, PowerPoint, or notes), keep the sections you want covered, pick a length, language, and voice, and press Create AI Video Lecture. Scholarly writes the narration from the pages you selected and renders an animated, chaptered video with synced captions, a searchable transcript, and question checkpoints. The Free plan includes one AI Creation Credit for life, enough for one video lecture of about 2 minutes. The catch: a free lecture is fixed at about 2 minutes, reads up to 32 pages, renders at 720p with the Scholarly logo, and can be watched and shared by link but not downloaded. Premium and Laureate lectures run up to 45 minutes, render at up to 1080p, can drop the logo, and download as an MP4.

It is built for one person learning from their own material: a textbook chapter before an exam, a certification syllabus on a commute, a dense report you need to understand by Monday. The video is narrated visuals made from your file, not a recording of you or a presenter.

What you get

  • A narrated, animated video scripted from your file, with one narrator from start to finish.
  • Chapters, one for each section you kept, so you can jump straight to the part you need.
  • Synced captions in the lecture's language, written from the same script as the narration, and a searchable transcript. Captions are included on every plan.
  • Interactive question checkpoints that slide up during playback, plus an AI chat beside the video that answers from the transcript and your original files.
  • On a Standard lecture, a companion slide deck.
  • On a paid plan, an MP4 download and an export pack: one zip with the video, the slide deck (Standard lectures), the narration script by chapter, captions as .vtt and .srt, and a README listing the sources.

What you can turn into a video

  • PDFs: textbooks, lecture handouts, papers, reports, manuals, scanned pages. Password-protected PDFs work; Scholarly asks for the password during upload.
  • Notes: text, Markdown, or rich-text files, or photos of handwritten notes.
  • Slides: PowerPoint (.pptx, .ppt) decks. A deck with sparse bullet points makes a thinner lecture than a full handout, so add the handout or speaker notes as a second source when you have them.
  • Word documents: .docx and .doc.
  • Also: images, video files, Google Drive files, a link to a website or an online PDF, or a typed topic on the Prompt AI tab with no file at all.

Step by step

1. Open the video lecture creator

Open New in the sidebar and choose Video lecture ("Narrated and rendered."). The Create AI Video Lecture window opens. You can also start from a file you just uploaded by choosing Video Lecture from its follow-up actions, or ask AI Chat: "Turn this PDF into a video lecture."

2. Add your source

Pick a tab:

  • Upload to drop in a file from your computer.
  • Library to use a file you already uploaded.
  • Google Drive to pick a file straight from your Drive.
  • Prompt AI to describe a topic instead.

You can also paste a link to a website or an online PDF. One lecture can combine 1 source on Free, up to 10 on Premium, and up to 20 on Laureate, for example a lecture PDF plus the matching textbook chapter. Then press Next.

3. Choose what to cover

Under What to cover, Scholarly finds the sections of a PDF and lists them with their page numbers. Keep the sections the lecture should teach; each one you keep becomes its own chapter. If sections cannot be found, for example in a scanned file with no text, choose Pick a page range instead and set the pages yourself.

Sections that would run past your plan's page limit are greyed out. The panel also tells you when your length and your pages do not match, for example that a short target makes the lecture an overview, or that the pages hold less explanation than the length you picked and a shorter length would avoid repetition.

4. Pick a design

Under Design, choose how the lecture teaches. Every design is available on every plan, and a new lecture starts on Standard.

DesignBest for
StandardA calm, editorial lecture that walks through the material. Includes the companion slide deck.
IllustratedA hand-drawn doodle explainer for concepts that benefit from pictures.
ShortA fast, vertical, phone-friendly recap of one thread. It has no length control.
Manim Math AnimationPrecise animated derivations for maths, physics, and engineering.
FreestyleThe AI reads your source and picks the style and tools.

See Video Modes for examples of each.

5. Choose the settings

SettingWhat it does
LanguageThe narration, captions, title, and chapter headings. Pick from over 70 options, or Auto to match your source.
Narration voiceLeave it on Let Scholarly choose, or open the picker, search by name or tone, and hear a short demo before you pick.
Use my script as writtenOff by default. Turn it on to have your source narrated word for word instead of a new teaching script.
Length (up to)Free lectures are about 2 minutes. Paid plans start on 15 minutes and the slider goes from 1 to 45 minutes.
Resolution720p by default; 1080p (Full HD) on paid plans.
LogoOn by default. Paid plans can turn it off for a watermark-free video. A logo cannot be removed later.
AI ModelAuto uses GPT 6 Luna for 1 credit. Models with a lock badge need a paid plan.

On desktop, Advanced options holds the extras: Background music, and on a Standard lecture a Look and Dark mode. Add custom instructions at the bottom, for example "Only cover Chapter 3", "Explain it for a beginner", or "Emphasize the key formulas".

When to use "Use my script as written". Use it when your file already is the script: lecture notes you wrote to be read aloud, a speech, or wording you must not change. The length follows your script within the length you choose, so a long script may be shortened to fit a short target; pick a longer target for a longer script. On Free, the 2-minute target means a longer script is cut to fit.

Not sure how to set it up? Plan with Chat, also under Advanced options on desktop, asks a few questions, then builds the video to your specifications.

6. Create

The Create AI Video Lecture button shows the total cost in AI Creation Credits before you start, counting the model and the length. Generation runs in the background, so you can close the window; a live progress page shows how many scenes are planned and recorded, and you can ask for an email when it is ready. A 2-minute lecture usually takes about 8 to 20 minutes, and a 45-minute lecture about 55 to 95 minutes. On a paid plan, the Length (up to) row shows the usual wait for the length you picked.

If a creation fails, its AI Creation Credit is returned automatically and the failure screen offers the fix that applies, such as picking fewer pages or replacing the source. See When a Creation Fails.

7. Watch, check, and download

  • Watch with chapters, captions (toggle them in the control bar), playback speed, and picture-in-picture. A lecture reopens where you stopped.
  • Check the lecture against your source. Under the player, the page shows the length you asked for beside the runtime delivered, the pages it was built from, and its language. The transcript is word for word what the narrator says, and the video's AI chat can tell you where a claim came from.
  • Download on a paid plan: open the three-dot menu and choose Download video for the MP4, or Download export pack for the video, slides, script, and captions in one zip. Captions stay a separate track and are not burned into the MP4, except on a Short, whose subtitles are part of the picture.
  • Share a link on any plan. People can watch, read the transcript, and browse chapters without an account.

To turn the lecture into practice, use Make next on the finished video page to create flashcards, a podcast, or a quiz from the lecture.

On Free, a finished lecture offers to make it longer. After you upgrade, Make longer remakes that lecture at about 10 minutes in the same place without using another credit, and the short version stays playable until the longer one is ready.

Plan limits

FreePremiumLaureate
Video lengthAbout 2 minutes, fixed1 to 45 minutes, 15 by default1 to 45 minutes, 15 by default
File size per upload8 MB300 MB1 GB
Uploads1 for lifeUnlimitedUnlimited
PDF pages per lectureUp to 32Up to 1,000Up to 3,000
Sources per lecture1Up to 10Up to 20
Resolution720p720p or 1080p720p or 1080p
Scholarly logoAlways onCan be turned offCan be turned off
Captions and transcriptYesYesYes
MP4 download and export packNo (view link only)YesYes
AI Creation Credits1 for life15 each week80 each week

What one lecture costs. The model's credits plus the length. On Auto (GPT 6 Luna) a lecture costs 1 credit. A target over 15 minutes adds 1 credit, and a target over 30 minutes adds 2, so on Auto a 15-minute lecture costs 1 credit, a 20-minute lecture 2, and a 45-minute lecture 3. Most other models also cost 1 credit before the length credits. The exceptions: Kimi K3 and Claude Sonnet 5.5 cost 2; GPT 6.1 Sol, GPT 6 Astra, and Claude Opus 5.5 cost 3; Gemini 4 Argon and Muse Wonder cost 4. If the finished video lands in a shorter length tier, the unused length credits are returned.

Weekly credits reset every Monday at 00:00 UTC. Credits are shared with every other AI creation, such as podcasts, slides, and flashcard decks. If you run out mid-week on a paid plan, you can buy an AI Credit Pack. See Plans and Limits.

Languages

Narration, on-screen text, captions, title, and chapter headings follow the language you pick, from over 70 options including Spanish, Portuguese, French, German, Arabic, Hindi, Japanese, Korean, and Chinese, with regional variants such as Brazilian Portuguese and Mexican Spanish. Your source and your video do not have to share a language: an English textbook chapter can become a lecture in Spanish. Arabic, Hebrew, and other right-to-left lectures are laid out right to left. With Use my script as written on, the script's own language sets the voice, title, and chapter headings. Which voices you can pick depends on the language. See Languages in Scholarly and Narration Voices.

Tips for a good result

  • Use a text PDF when you can. Scanned and photographed pages work, and Scholarly reads diagrams, tables, and handwriting, but accuracy follows the scan quality.
  • Narrow the pages. One chapter makes a sharper lecture than a whole textbook. Make one video per chapter and keep them in a project.
  • Match the length to the material. A few pages hold a few minutes of explanation. If the panel says your pages would start repeating at your target, take the length it suggests.
  • Pair thin slides with the handout. Bullet-point decks leave the AI guessing. On a paid plan, add the handout, the textbook pages, or the speaker notes as a second source.
  • Say who it is for. A custom instruction such as "I'm studying for the CPA exam, focus on the tested rules" or "Explain like I'm new to this" changes what gets emphasized.
  • Pick Manim Math Animation for derivations and Illustrated for concepts that need pictures.
  • Turn the logo off before you generate if you need a clean MP4. It cannot be removed afterwards.

What it does not do

  • No avatar or on-screen presenter. The video is narrated visuals: definitions, comparisons, step sequences, charts, equations, and illustrations.
  • No recording of your own voice. Narration uses Scholarly's AI voices.
  • No timeline or scene-by-scene editor. To change what the video says, adjust the pages, the source, or the custom instructions and generate again. You can rename a finished video.
  • No download on Free. Free lectures can be watched and shared by link only.
  • No lecture longer than 45 minutes. For a long book, make one lecture per chapter.
  • No guarantee it is error-free. The narration is written from your pages, not the open web, but check the transcript before you rely on it for an exam.

Facts at a glance

  • Inputs: PDF, Word (.docx, .doc), PowerPoint (.pptx, .ppt), text and Markdown notes, images, video files, Google Drive files, website or PDF links, or a typed topic.
  • Output: narrated, animated video with chapters, synced captions, a searchable transcript, question checkpoints, AI chat, and (Standard design) a companion slide deck.
  • Length: about 2 minutes on Free; 1 to 45 minutes on Premium and Laureate.
  • Pages per lecture: up to 32 on Free, 1,000 on Premium, 3,000 on Laureate.
  • File size: 8 MB on Free, 300 MB on Premium, 1 GB on Laureate.
  • Languages: over 70 options, right-to-left layout for Arabic and Hebrew.
  • Downloads: MP4 and export pack (video, slides, script, .vtt and .srt captions) on paid plans; a view link on every plan.
  • Cost: 1 AI Creation Credit on Auto up to 15 minutes; +1 above 15 minutes, +2 above 30 minutes.

Frequently asked questions

Can I turn a whole textbook into one video?

You can select it on a paid plan, but one video per chapter teaches better. Use What to cover to keep one chapter at a time.

Will the video say things that are not in my PDF?

The narration is written from the pages you selected. Read the transcript against the source anyway, and turn on Use my script as written if the wording must stay exactly as written.

Can I use my PowerPoint slides?

Yes. Upload the .pptx or .ppt file. If your slides are mostly bullet points, add your notes or the handout as a second source on a paid plan so the narration has more to work from.

Can I download the video?

On Premium and Laureate, yes: Download video saves the MP4, and Download export pack adds the slides, script, and captions. On Free you can watch and share a link.

Are captions included?

Yes, on every plan, in the language of the lecture. On paid plans the export pack includes them as .vtt and .srt files; a Short has its subtitles burned into the video instead.

Can I make the lecture in another language than my PDF?

Yes. Set Language to the language you want. The source can stay as it is.

How long does it take?

It depends on the length. A 2-minute lecture usually takes about 8 to 20 minutes, a 45-minute lecture about 55 to 95. You can close the window while it works.

Was this helpful?