Start free

Blog · · 4 min read

Professor lecture to syllabus: turning recorded lectures into study guides

Recorded lectures are unsearchable by default. How professors turn 90-min talks into syllabus updates, study guides, and FAQ docs students actually use.

.5 Sonnet handles 200,000 tokens — roughly 15–20 hours of lecture audio in transcript form. Gemini 1.5 Pro handles up to 2 million tokens, roughly 1.5 million words or 150+ hours of lecture transcripts, which fits an entire semester of a course. The practical implication: you can ask "across all 24 lectures, which concepts did I cover but never define formally?" and get an answer. That question is impossible to ask of a stack of MP4s.

What the AI cannot do

It cannot tell you which 30% of your lecture was the important part. It cannot identify the diagram you drew on the whiteboard, and it does not OCR your slides. It cannot grade student understanding. It should not invent exam emphasis you never voiced — if you said "this distinction matters", the guide can flag it; if you didn't, the model should not pretend you did. The output will be comprehensive and slightly flat — a competent TA's first draft, not a senior professor's final cut. Plan to edit.

Accessibility: the April 2026 ADA deadline is not optional

The U.S. Department of Justice finalized a rule in April 2024 requiring public universities to bring all web content — including lecture audio and video — up to WCAG 2.1 Level AA. The deadlines:

  • April 24, 2026 for public entities in jurisdictions with populations over 50,000.
  • April 24, 2027 for smaller jurisdictions.

WCAG 2.1 Level AA includes synchronized captions for prerecorded video Success Criterion 1.2.2) and a transcript for prerecorded audio (1.2.1). If your university posts lecture recordings to a course portal and you teach at a public institution, this applies to you.

The cost math is what forces the decision. Human captioning runs around $1.99 per minute per Rev.com's published 2026 rates. A 90-minute lecture is $179. A 14-week course with two lectures a week is roughly $5,000. A department with 200 lecture courses is looking at six figures a year. AI transcription brings the marginal cost to roughly $0.36–$0.37 per hour at the API layer, or 1 credit per minute on our Pro plan — 600 audio-minutes/month, $19/month as of May 2026.

The honest caveat: AI captions are not legally equivalent to verified human captions in every accommodation context. A student with a documented hearing accommodation may still require human-verified captions for their specific course. The right read is that AI can get you to a WCAG-compliant baseline for the whole catalog when the captions are reviewed for accuracy, and you reserve full human captioning for the accommodations that demand it.

We don't ship an ADA compliance certification — no transcription engine does. What we ship is the transcript and a caption file (SRT/VTT) that your LMS team can attach to the video. All 99 languages we support are priced the same — one price, every language — so a Spanish literature seminar costs the same per minute as an undergraduate stats lecture.

Try it on your audio

Start free →

30 minutes a month, no card.

The weekly workflow that actually sticks

Professors who make this stick don't batch at semester end. They run a 15-to-30-minute-per-lecture Friday loop:

  1. Pull the week's recordings off Zoom Cloud or the lecture-capture system.
  2. Upload to our video-to-text pipeline — most lecture-capture systems output MP4, so this is the right entrypoint.
  3. Skim each transcript for the four extraction targets above. For a clean 60–90 minute lecture, budget 15–30 minutes of review. Technical vocabulary, multiple un-mic'd speakers, or accessibility publication push that higher.
  4. Update the syllabus document. Post the study guide to the LMS. Add the SRT to the recording.
  5. File the transcript into a single course archive with a consistent name — BIO201_week04_cell_signaling_2026-02-12 — and topic tags.

Step 5 is the one that compounds. After two semesters, the archive is searchable across every lecture you've ever given in that course. You can answer "when did I last explain the bootstrap?" in three seconds. You can find the version of an explanation that worked best for students. You can hand the archive to a co-instructor.

Otter.ai and Rev.com handle steps 1–2 well — Otter is strong on live capture inside a meeting, Rev is strong when you need human-verified output, Descript shines if you're editing the lecture as a video. The seam most of our academic users hit with general-purpose tools is the archive layer: tagging, cross-lecture search, and exporting clean caption files at scale. That's where a pipeline built around lecture notes specifically tends to fit better.

What we don't ship for this workflow

To be clear about the seams:

  • Live in-room captioning: we transcribe the recording after the lecture ends. We don't project live captions on screen. If you need that, look at Microsoft Teams or Google Meet's built-in live captions, or a CART provider for accommodated students.
  • LMS integration: no native Canvas/Blackboard/Moodle plugin yet. You download the transcript and upload it to your LMS, or wire a webhook.
  • Slide / whiteboard capture: audio only. Slides are not OCR'd, the whiteboard photo isn't part of the transcript.
  • HIPAA BAA: we do not sign BAAs yet. If you teach a clinical course and your lectures include real, identifiable patient data, don't upload those recordings here.
  • Native mobile app: web only, mobile-responsive. You can upload from a phone browser.

One more privacy note: lectures contain student voices, sometimes with personal disclosures, accommodation discussions, or named third parties. Review before publishing transcripts to a public LMS page.

What next

  • Pick one course you'll teach next semester. Record one 60-minute lecture this week and run it through a Free plan upload — 30 minutes/month, exports unlocked, no card required.
  • Read the transcript with the syllabus open. Mark every place the lecture and the syllabus disagree. That list is your winter break to-do.
  • If you teach at a public U.S. university, check with your accessibility office about which deadline applies — April 2026 or April 2027 — and whether they want SRT or VTT caption files.
  • If you're running an entire department's catalog through this, email us before you upload 5,000 hours. We'll size the Business plan and the overage packs to match.