All plays › Coaches

CoachesClaude + Claude Code (with Auphonic, ffmpeg)

Turn messy screen recordings into finished course lessons in minutes

Result

self-reported

Replaced four contractor roles ($300–500/month sound, $25/hour privacy blur, $75–100 per video edit, $40 per graphic). Turnaround went from 4 days to 8 minutes per lesson.

Aryan Mahajan’s team records raw 30-minute walkthroughs and runs one command. Auphonic cleans the audio, Claude finds and blurs client names and emails frame by frame, cuts false starts from the transcript, and Claude Code builds thumbnails and guides from saved templates before the lesson goes live.

Who did it

Aryan Mahajan (@aryanXmahajan)Zoro Corp (Beverly Hills)Aryan Mahajan on X · Oct 16, 2026

Try it now

  1. Record your lesson in one take. Don’t stop for mistakes.
  2. Run the audio through Auphonic (or similar) to clean echo and set loudness.
  3. Have Claude read the transcript timestamps and list every false start and long pause to cut, then cut with ffmpeg.
  4. Have it scan for client names, emails and open tabs to blur, and re-check the final export before publishing.
Copy-paste prompt
Here is the word-level transcript with timestamps for my lesson [file]. List every false start, repeated sentence and pause over 1.2 seconds as cut ranges (start–end). Then write the ffmpeg command to remove them. Separately, list every moment a client name, email address or private window may be on screen so I can blur it, and tell me how to re-check the final file.

Want this set up on your own site?

Build it with Page

What to watch out for

Check the privacy blur yourself before anything goes public; one missed email is a real problem. His company sells AI builds, and the savings are his own figures. Build it with Page: a course page where each finished lesson lands.

Did it work?

PageMotor keeps this site free.