AI Video Processor — Batch Extract Transcript, Summary, and Screenshots
Upload an MP4, MOV, WebM, or AVI video and AI transcribes the audio, generates a summary, and extracts scene-change screenshots as a ZIP — all at once. Ideal for turning Zoom / Teams recordings, YouTube videos, lecture recordings, and internal training videos into meeting notes or summaries.
Three outputs from one video
This tool automatically generates the following from a single video file:
1. **Full transcript**: Japanese audio transcribed with high accuracy by OpenAI Whisper 2. **Summary (Markdown)**: Claude Sonnet reads the transcript and extracts key points 3. **Screenshot ZIP**: screenshots captured on scene change or at fixed intervals
A one-hour meeting video is fully processed in 5–10 minutes.
Use cases
· **Meeting minutes from Zoom / Teams recordings**: record → upload → structured minutes in 5 minutes · **Lecture and seminar notes**: transcript + chapter screenshots from video lectures become complete notes · **Article from a YouTube video**: transcribe your own video → AI-format into an article · **Training material review**: watch a training video on the Free plan → minutes + key-scene archive
Technology stack
· **Audio extraction**: ffmpeg (LGPL license, commercially OK) cascades video to mp3/m4a/wav · **Transcription**: OpenAI Whisper API (`whisper-1`, language=ja for Japanese) · **Summarization**: Claude Sonnet 4.5 · **Screenshots**: ffmpeg scene change detection or fixed-interval captures
Privacy
Uploaded videos are stored temporarily in AWS S3 (Tokyo region) and automatically deleted within one hour. Whisper API and Anthropic API transfers are TLS-encrypted + zero-data-retention contracts. Safe for confidential internal meeting recordings.
File size: Free 20 MB / Light 100 MB / Premium 200 MB. Video processing runs as an async job — if you close the browser, processing continues.