AI Lecture Summary — From Recording to Study Notes

VexaScribe generates a structured lecture summary with: an executive overview, timestamped chapter breakdown, key concepts list, terminology with definitions, examples given, and takeaways. All sections are generated from the audio — you don't need to paste any text.

Generic summarizers work on text you paste in. VexaScribe takes your audio or video file directly, transcribes it first, then runs a lecture-specific summary that extracts chapters with timestamps, terminology, and examples — structure that generic tools miss.

A 1-hour lecture transcribes in about 5-10 minutes, then summarization adds another 1-2 minutes. Total: roughly 10-12 minutes for a full transcript plus structured summary.

Yes. Upload MP4, MOV, MKV, or other video formats. VexaScribe extracts the audio automatically and transcribes it, then summarizes. Works with Zoom recordings, downloaded course videos, and YouTube lectures saved locally.

Yes. The chapter breakdown in the summary includes timestamps linked to the original recording. You can click any chapter to jump to that moment in the audio player.

Summarization is included with every plan at no extra charge. You pay only for transcription minutes. Plans start at $2/month for 200 minutes (~3.3 hours of lectures). A typical semester of weekly 1-hour lectures (30 hours) costs about $6–$18 depending on plan.

VexaScribe supports 99 languages for transcription. Summaries are generated in the same language as the lecture — no translation needed.

Yes. Export the summary as plain text, or download the full transcript with timestamps in TXT, SRT, or VTT format. Summary and transcript are both included in every export.

Yes. The DOCX export imports cleanly into OneNote. For Notion, copy-paste from the transcript editor preserves headings, bullet lists, and speaker labels. For Obsidian, export as TXT and paste into a Markdown file — VexaScribe's structured output (Executive Summary → Chapters with timestamps → Key Concepts → Terminology → Takeaways) maps naturally to nested Markdown headings. For Anki flashcard export, we don't do that directly — pipe the Key Concepts + Terminology section into an AI flashcard generator (NoteGPT, RemNote) as a second step.

Depends on your study method. Bullet summary is fastest for quick review (~5% of lecture length). Cornell notes work best for active recall — 2-column question/answer with cue column, per Walter Pauk's original Cornell method. Concept maps work best for interconnected topics (organic chemistry pathways, historical timelines). Study guide format is best for exam prep — includes definitions, examples, and sample questions. VexaScribe generates all four from the same lecture recording.

Lecture transcription is the full verbatim text of everything the professor said — typically 8,000-12,000 words for a 60-minute lecture — used for quote extraction, exam-question wording verification, or accessibility. Lecture summary is structured study notes (executive summary + chapters + key concepts) — typically 500-1,500 words — used for quick review and exam prep. VexaScribe produces both from one upload. See /lecture-transcription for the full transcript workflow.

Summary quality depends on transcription quality first. Clean audio (professor with lapel mic, Zoom recording, laptop mic in quiet room) transcribes at 92-97% accuracy per the Whisper Large-v3 benchmarks — the summary generated from that transcript captures all major topics reliably. Poor audio (open-mic in noisy lecture hall, phone in your bag) drops transcription to 78-85% — the summary will miss or misstate some concepts. Best practice: sit near the professor, use lapel mic if allowed, or record straight from Zoom.

Yes. Export the Zoom Cloud Recording as MP4 and upload directly to VexaScribe — audio is extracted automatically. If you want native Zoom transcription instead, Zoom AI Companion is included on Pro plans and above. For post-lecture summary workflow with better speaker labels and export to Notion/OneNote/Obsidian, upload the Zoom MP4 to VexaScribe.