Split Long Class Recordings for AI Transcription

The transcription limit for a single recording on GPAce is 2 hours. This keeps AI transcription stable across providers and controls processing cost. For longer audio, split it by chapter and upload each segment separately.

If you record a full - day workshop, a Lab Final Review, or a 4 - hour consecutive lecture into one file and upload it to GPAce, it will be directly rejected.

The good news is that you don't need to install new software or learn commands to split the audio. You can choose any of the following four methods and finish it in 5 minutes.

💡 After splitting, upload each segment independently under the same course. GPAce will generate independent notes, quizzes, and flashcards for each segment, making it easier to focus on different chapters during review.

Method 1: macOS Built - in GarageBand (Simplest, No Installation Required)

GarageBand is a free music editor built into macOS and iOS. It's more straightforward for audio editing than iMovie.

Open GarageBand → Create a new project → Select "Empty Project" → Press Cmd + N to create a new track and choose "Audio" as the type.
Drag your class recording in mp3 or m4a format into the GarageBand timeline.
Find the position of 2:00:00 at the top of the timeline → Press Cmd + T to cut at the playback head position.
Select the second segment → Press Delete to delete it. Then, go to the menu Share → Export Song to Disk → Select mp3 → Name it lecture - part1.mp3 → Save.
Undo the deletion (Press Cmd + Z twice) → Delete the first segment and keep the second half. Then, export it as lecture - part2.mp3 in the same way.

Each segment is an independent mp3 file. You can upload them to the same course on GPAce in sequence.

Method 2: Audacity (Available on Windows, macOS, and Linux, Free and Open - Source)

Audacity is an established audio editing tool, completely free and ad - free.

Download Audacity → Install it → Open it.
Go to File → Open and select your class recording.
There is a time scale below the timeline. Drag the mouse to select the segment from 0:00:00 to 1:55:00 (Leave a 5 - minute buffer to avoid errors).
Go to File → Export → Export Selected Audio → MP3, and name it lecture - part1.mp3.
Select the segment from 1:55:00 to the end → Export it in the same way and name it lecture - part2.mp3.

Audacity's splitting is based on "exporting selected areas", and the original file will not be modified. You can operate repeatedly without worry.

Method 3: ffmpeg Command Line (Fastest, Can Be Done in 10 Seconds, One - Time Installation Required)

If you are already using the command line (macOS Terminal / iTerm / Windows PowerShell), ffmpeg is the fastest way to split audio.

One - Time Installation

# macOS
brew install ffmpeg

# Windows (PowerShell, requires winget)
winget install Gyan.FFmpeg

# Linux (Ubuntu/Debian)
sudo apt install ffmpeg

Split into Two Segments

Assume the original file lecture - 2025 - 05 - 24.mp3 is 3 hours long:

# First 2 hours (7200 seconds)
ffmpeg -i lecture - 2025 - 05 - 24.mp3 -t 7200 -c copy lecture - part1.mp3

# From 2 hours to the end
ffmpeg -i lecture - 2025 - 05 - 24.mp3 -ss 7200 -c copy lecture - part2.mp3

-c copy means no re - encoding, which can be completed in seconds without loss of audio quality.

Split into N Segments (Automatically Split Every 110 Minutes)

ffmpeg -i lecture - 2025 - 05 - 24.mp3 \
       -f segment -segment_time 6600 -c copy \
       lecture - part%02d.mp3

It will generate lecture - part00.mp3, lecture - part01.mp3, lecture - part02.mp3... Each segment is about 110 minutes long.

Method 4: Online Tools (Don't Want to Install Any Software)

If you only need to split audio occasionally, browser - based online tools are sufficient. However, when the recording contains class content, it is recommended to process it locally (for privacy, network speed, and the single - file size is usually ≤ 100 MB).

Tool	Features
mp3cut.net	Chinese - friendly; Pure local JavaScript processing, no upload to the server
audio - trimmer.com	Also local processing, clean UI
clideo.com/cut - audio	Supports m4a / wav; Free version has no watermark? Yes, generally no audio watermark

The operations are similar: Drag the file → Drag the timeline to select the area → Export.

Best Practices After Splitting

Name the files with part numbers: CS101 - Lec - 12 - part1.mp3 / CS101 - Lec - 12 - part2.mp3. It will be clear when the recordings are sorted by time in the GPAce list.
Upload them to the same course in sequence: GPAce will generate independent notes for each segment. Read them in chronological order during review.
View the mind maps in segments: A separate mind map for each segment is more focused than a single 3 - hour map.
Accumulate flashcards: All flashcards from each part will enter the review pool for this course, and the Ebbinghaus reminders will be merged across segments.
If the cut point is in the middle of a key sentence: Use Method 2 (Audacity) to listen around the cut point and move it to a natural pause to avoid missing one or two terms in the notes.

Frequently Asked Questions

Q: Why not directly extend the limit to 5 hours?
A: Very long recordings significantly increase transcription failure risk and processing cost, and they also make the generated notes harder to review. Splitting recordings into segments under 2 hours keeps both reliability and note quality more stable.

Q: Will splitting the audio reduce the quality of the notes?
A: Almost not. At most, one or two words may be lost near the cut point. If you use Audacity to cut at a natural pause, there will be no impact at all. On the contrary, each segment generates independent notes, making the chapter boundaries clearer and preventing you from getting lost in a 3 - hour long note during review.

Q: Will GPAce automatically merge multiple parts into one large note?
A: Currently, it will not merge automatically. Each segment has its own note and a set of flashcards. If you need an "overall semester note", it is recommended to export the Markdown of each segment and manually splice them (you can click "Copy Markdown" in the upper - right corner of each note).

Q: I record classes directly in the browser and the recording exceeds 2 hours. What should I do?
A: GPAce's "online recording" function (the Tab of /recordings/upload/<course_id>) has no time limit, but the browser's memory is limited. It is recommended to "stop and upload" after each class instead of recording the whole semester at once.

Record a class → Split the audio → Upload → Get a complete review package of bilingual notes, quizzes, and flashcards in a few minutes. Wish you an efficient review and a smooth final exam.