Extract audio with ffmpeg for transcription One command, no re-encoding.
Run ffmpeg -i video.mp4 -vn -c:a copy audio.m4a: it lifts the audio track out of the video without re-encoding, so a 3-hour recording at 128 kbps becomes a file of about 173 MB instead of several gigabytes. WhipScribe takes that file, or the original video, up to 10 hours or 5 GB, and transcribes it at about 2 minutes per audio hour with the first 20 minutes free.
ffmpeg for first-timers
Two commands, then one upload.
Open a terminal and install ffmpeg once
On a Mac, press Command and Space, type Terminal and press Return. On Windows, right-click Start and pick Terminal or Windows PowerShell. Install ffmpeg one time with brew install ffmpeg on a Mac that has Homebrew, winget install Gyan.FFmpeg on Windows, or sudo apt install ffmpeg on Ubuntu. Then type cd and a space, drag the video's folder into the window and press Return.
Command one copies the sound as it is
Type ffmpeg -i lecture.mp4 -vn -c:a copy lecture.m4a and press Return. The -vn part drops the picture and -c:a copy keeps the audio bytes exactly as recorded, so nothing is lost and the job is quick. Most MP4 files from phones, Zoom and screen recorders carry AAC audio, which belongs in an .m4a file. If ffmpeg stops with an error about the codec or container, use the second command.
Command two makes a small MP3
Type ffmpeg -i lecture.mp4 -vn -ac 1 -c:a libmp3lame -b:a 64k lecture.mp3 instead. It encodes the speech again as mono at 64 kbps, which is 0.48 MB per minute: about 29 MB an hour, 86 MB for three hours and 288 MB for ten. Speech stays clear at that rate, though the encode takes longer than a straight copy.
Then send the small file
The audio file uploads in a fraction of the time the video needs. Drop it in the box above, or go to the upload page for anything over 500 MB. About 2 minutes per audio hour later you have the transcript plus a brief with thesis and takeaways, chapters with timestamps, and the names, numbers and dates that were said.
Video or audio
Why sending only the sound is quicker.
✗ Uploading the whole video
A long video is mostly picture. The words sit in a track that is a small share of the file, and sending everything else costs time and sometimes fails halfway.
- A 5 Mbps video is 2.25 GB per hour
- Uploads stall on hotel or campus Wi-Fi
- Over 2 GB waits until a pack is on the account
- Over 5 GB will not go up as one file
- Web converters that redo the picture too
- A browser tab that must stay open for an hour
✓ Sending the audio track
ffmpeg copies the track out, and the transcript is made from exactly the same sound.
- Audio copied with no change in quality
- 128 kbps AAC at 57.6 MB per hour
- 64 kbps mono MP3 at 28.8 MB per hour
- Ten hours of audio well under 1 GB
- Text at roughly 2 minutes per audio hour
- The first 20 minutes free
Sample output
Text from a track pulled out of a lecture video.
Illustrative excerpt, not a real class. The timestamps match the original video, because copying the track leaves its clock alone.
Export
One transcript. Five clean formats.
Every pack exports all five — no format is held back for a higher tier.
Plain text
De-ummed paragraphs. Ready to paste.
SRT captions
Timed to the speech. Every video editor reads this.
WebVTT
HTML5 player + YouTube uploads.
Show notes
Formatted with chapters and pull-quotes.
Machine-readable
Timestamped text for your own tools.
Pricing
Honest pricing, no surprises.
A pack from $6 (500 minutes) opens the full transcript. After that, no subscription and no auto-renew — every tier is a one-time minute pack, and credits never expire. Buy once, use the minutes whenever you need them.
Casual pack
$6/500 min
8 hours of audio — $0.72 an hour. One-time — credits never expire.
- 500 minutes of audio
- All export formats
- Unlimited AI chat per transcript
- Executive summary & key quote
Best value
$12/2,000 min
33 hours of audio — $0.36 an hour. Credits never expire.
- 2,000 minutes of audio
- Priority queue
- All export formats
- Unlimited AI chat per transcript
Bulk pack
$24/5,000 min
83 hours of audio — $0.29 an hour. Credits never expire.
- 5,000 minutes of audio
- Priority queue
- API access for batch pipelines
- All export formats
FAQ
ffmpeg and transcription, answered.
How long do the extraction and the transcript take?
The copy command reads the video and writes the audio without encoding anything, so on a laptop with an SSD it usually finishes in well under a minute, even for a multi-hour file. The MP3 command encodes every second and often takes a few minutes. Transcription then runs at about 2 minutes per audio hour, so a 3-hour lecture is text roughly 6 minutes after the upload ends.
What does a 3-hour recording cost once the audio is out?
Three hours is 180 minutes. A first-time visitor gets the first 20 minutes free, and the $6 pack (500 minutes) transcribes the rest and leaves 320 minutes on the account. Pulling the audio out changes the upload size, never the minutes counted, because the length of the talk is the same either way.
How big will the audio file be, and what if the upload refuses it?
Size is bitrate times time. AAC at 128 kbps is 16 KB a second, which is 57.6 MB an hour; MP3 at 64 kbps mono is 28.8 MB an hour. Ten hours at either rate stays under 1 GB, far inside the 2 GB that applies before the first pack. If the upload refuses the file, convert it to MP3 first with the second command; that covers unusual codecs inside MKV or MOV files.
Do I need to cut a long file into parts?
Only past 10 hours or 5 GB. Beyond 10 hours, ffmpeg can cut the audio in two without encoding: ffmpeg -i long.m4a -t 05:00:00 -c copy part1.m4a makes the first five hours, and ffmpeg -ss 05:00:00 -i long.m4a -c copy part2.m4a makes the rest. Upload them as two files.
Does the language of the video change the command?
No. The command is the same for a Korean seminar, a Portuguese podcast or a Tagalog class, and the language is detected automatically among 99 on our side. A recording that moves between English and Tagalog is handled. Cebuano, Ilocano and the other regional Philippine languages are supported, but accuracy trails the major languages, so treat that text as a fast first draft and plan one review pass.
Which formats can I download, and where is the extracted audio processed?
Five downloads come back from an extracted track: TXT, DOCX, SRT, VTT and JSON. A copied track usually starts at the same zero as the video, so SRT or VTT captions made from the audio drop back onto the original file. Extracted audio is transcribed on our own servers and is never used for training.
Related
More on formats and file size.
Sound out of the video. Text about 2 minutes per hour later.
Upload the audio fileOperated by Neugence Technology Pvt. Ltd. · contact@whipscribe.com · Security · Privacy · Terms