How to split an audio file into multiple parts
- Step 1Open the splitter and drop your file — Go to the audio splitter and drop one audio file (MP3, WAV, FLAC, M4A, OGG, or Opus). The file is read into memory in your browser — it is never sent to a server. The splitter takes a single file at a time.
- Step 2Leave Split mode on "Every N seconds" — The Split mode dropdown defaults to Every N seconds, which is what equal-parts splitting uses. The other two values — On silence and Manual markers — are covered in the sibling guides; for fixed-length chunks, keep the default.
- Step 3Set the Segment (s) length — Enter your chunk length in the Segment (s) field. It accepts 5 to 3600 seconds in steps of 5. Example: enter
600for ten-minute parts, or300for five-minute parts. The tool computes the number of parts as ceil(duration ÷ segment). - Step 4Run the split — Start processing. FFmpeg cuts each part with
-ss <start> -t <length>and re-encodes it with the encoder for your input format. A part shorter than 0.05 s (a rounding sliver at the very end) is dropped automatically. - Step 5Review the part list — The result panel shows the part count and each part's size. Confirm the count matches your expectation — for a 25-minute file at
300seconds you should see 5 parts (four full + one 5-minute remainder, or five exact parts if the duration divides evenly). - Step 6Download every part — Click Download all to save each part as its own file, or use the per-row Download link to grab a single part. Files are named
<original-name>-part-1,<original-name>-part-2, and so on. There is no ZIP archive — each part is a standalone file.
Split modes and their controls
The splitter exposes exactly three modes via the Split mode dropdown. Defaults and ranges are read from the tool's controls. Equal-parts splitting uses the first row.
| Mode (dropdown value) | What it does | Control(s) | Default / range |
|---|---|---|---|
| Every N seconds | Cuts the file into back-to-back parts of a fixed length; final part holds the remainder | Segment (s) | 60 s · range 5–3600 s, step 5 |
| On silence | Detects silent regions and cuts at the midpoint of each one (best for natural pauses) | Threshold (dB), Min silence (s) | -40 dB (range -80 to -10) · 0.6 s (range 0.1–5) |
| Manual markers | Cuts at the exact second positions you add to the marker list | Add marker → list of seconds | empty list; markers sorted, only 0 < m < duration kept |
Equal-parts math: file duration vs Segment (s)
Number of output parts for common segment lengths. Parts count = ceil(duration ÷ segment). The last part is shorter when the duration does not divide evenly.
| File duration | Segment (s) = 300 (5 min) | Segment (s) = 600 (10 min) | Segment (s) = 900 (15 min) |
|---|---|---|---|
| 25:00 (1500 s) | 5 parts (all 5:00) | 3 parts (10, 10, 5:00) | 2 parts (15:00, 10:00) |
| 47:30 (2850 s) | 10 parts (last 2:30) | 5 parts (last 7:30) | 4 parts (last 2:30) |
| 1:30:00 (5400 s) | 18 parts (all 5:00) | 9 parts (all 10:00) | 6 parts (all 15:00) |
Tier limits that apply to the splitter
The splitter is a Pro-tier tool, so the Free row cannot run it — it is shown for context on file-size and duration ceilings across the audio family.
| Tier | Max file size | Max duration | Files per run | Runs splitter? |
|---|---|---|---|---|
| Free | 50 MB | 30 min | 1 | No (Pro-only tool) |
| Pro | 200 MB | 120 min | 10 | Yes |
| Pro-media | 100 GB | unlimited | 100 | Yes |
| Developer | 100 GB | unlimited | unlimited | Yes |
Cookbook
Concrete equal-parts recipes. Each shows the field you set and the parts you get. Marker values are in seconds; "part" files are separate downloads, not a ZIP.
Chop a 90-minute capture into ten-minute blocks
The everyday case: a long single file you want in even blocks for review or archiving. One field, predictable output.
Input: meeting.mp3 (1:30:00 = 5400 s) Mode: Every N seconds Segment (s): 600 Result: 9 parts, each 10:00 meeting-part-1 ... meeting-part-9 (all .mp3, re-encoded with libmp3lame) Download all -> 9 individual files saved
Five-minute parts with a short remainder tail
When the duration does not divide evenly, the final part is simply shorter — it is not padded and not dropped (unless under 0.05 s).
Input: lecture.wav (47:30 = 2850 s) Mode: Every N seconds Segment (s): 300 Parts count = ceil(2850 / 300) = 10 parts 1-9 : 5:00 each (2700 s) part 10 : 2:30 (the 150 s remainder) Output format: .wav (PCM 16-bit), same as input
Keep WAV lossless-as-source (no transcode to MP3)
The splitter has no format selector, so output always matches the input. A WAV in stays a WAV out — handy when a downstream transcription tool wants uncompressed audio.
Input: interview.wav (PCM 16-bit, 44.1 kHz) Mode: Every N seconds Segment (s): 1200 (20-minute parts) Each part is re-encoded with pcm_s16le (WAV). Note: this is a re-encode, not a byte-exact copy, but WAV->WAV PCM is mathematically lossless.
Pre-strip dead air, then split into even parts
Combine tools: tighten the recording first so the equal parts contain more speech per part, then split. Two separate tools, run in sequence.
Step 1: silence-stripper -> removes long pauses
(see /audio-tools/silence-stripper)
Step 2: splitter, Every N seconds, Segment (s): 600
Result: denser 10-minute parts (less dead air),
same equal-parts behaviour.Split, then re-merge a subset
If you split too finely, the merger rejoins parts end-to-end. Useful for trial-and-error editing without re-cutting from scratch.
splitter -> 18 parts of 5:00 each (pick parts 3, 4, 5) audio-merger (/audio-tools/audio-merger) -> drop the three parts in order -> one 15:00 file, concatenated
Edge cases and what actually happens
Final part is shorter than the segment length
By designWhen the file duration does not divide evenly by your Segment (s) value, the last part holds the remainder and is shorter than the rest. A 25:00 file at 600 seconds yields three parts: 10:00, 10:00, and a 5:00 tail. This is expected — the tool never pads the final part with silence.
A rounding sliver at the very end (< 0.05 s)
DroppedIf a computed part is shorter than 0.05 seconds — usually a floating-point rounding remnant when the duration is an exact multiple of the segment — that part is skipped rather than written as a near-empty file. You will not see an empty part-N in the output list.
Expecting a single ZIP of all parts
Not supportedThe splitter outputs separate files, not a ZIP archive. Download all saves each part as its own download in one click; there is no option to bundle them. If your browser blocks multiple sequential downloads, allow them for this site or use the per-row Download links.
Wanting the output in a different format than the input
Not supported hereUnlike the merger, the splitter has no format selector — output always matches the input extension and codec. To split and change format, convert first (for example with mp3-to-wav or wav-to-mp3) and then split the converted file.
Hoping for a lossless / stream-copy cut
Re-encodedEvery part is re-encoded (-c:a with the encoder for your format), so MP3 parts incur one generation of lossy re-encode at the cut boundaries. For a single lossless cut without re-encoding, use the audio trimmer, which has a stream-copy option. WAV and FLAC re-encodes remain mathematically lossless.
Free tier user tries to split
Tier blockedThe splitter requires a Pro tier (minTier: pro). On the Free tier the run is blocked. Pro allows up to 200 MB and 120 minutes per file; Pro-media and Developer remove the duration cap (100 GB file ceiling).
File exceeds your tier's size or duration limit
RejectedPro caps a single file at 200 MB and 120 minutes. A 3-hour podcast on Pro exceeds the duration limit and is rejected before processing — split it on a Pro-media or Developer plan (unlimited duration), or trim it first with the audio trimmer.
Segment value below the minimum
ClampedThe Segment (s) field accepts 5–3600 seconds. Values below 5 are not accepted by the control. If you need sub-5-second clips, use manual markers where each cut position is a precise second value (step 0.1).
Very large part count from a tiny segment
Expected (slow)Setting Segment (s) to 5 on a 90-minute file produces ~1080 parts, each re-encoded individually in WebAssembly. It will run, but slowly, and your browser will fire ~1080 downloads on Download all. Prefer larger segments unless you genuinely need micro-clips.
Frequently asked questions
How do I split an audio file into equal parts?
Drop the file, keep Split mode on Every N seconds, and set Segment (s) to the length you want each part to be (in seconds). The tool cuts back-to-back parts of that length and drops the leftover into a final shorter part. For example, 600 gives ten-minute parts.
Is it actually free?
The splitter is a Pro-tier tool — it requires a Pro plan or above to run, even though the page targets a 'free splitter' search. Pro covers files up to 200 MB and 120 minutes. If you only need a single lossless cut, the audio trimmer is available on the Free tier.
Does splitting upload my file anywhere?
No. All processing runs in your browser through FFmpeg 8.1 compiled to WebAssembly. The audio is read into memory locally and never transmitted to a server, so private recordings stay on your device.
Do I get a ZIP of all the parts?
No. Each part is a separate file. Download all saves every part as its own download in one click, and each row has a Download link if you only want one part. There is no ZIP bundling option.
What format are the output parts?
The same format as your input. The splitter has no format selector, so an MP3 in produces MP3 parts (libmp3lame), a WAV produces PCM 16-bit WAV parts, FLAC stays FLAC, and so on. Convert before splitting if you need a different output format.
Is the split lossless?
Every part is re-encoded, so MP3 parts pick up one generation of lossy re-encode. WAV and FLAC re-encodes are mathematically lossless. For a truly stream-copied lossless cut, use the audio trimmer's lossless option instead.
How many parts will I get?
ceil(duration ÷ Segment). A 47:30 file (2850 s) at 300 seconds gives ceil(2850/300) = 10 parts, the last one being a 2:30 remainder. The result panel shows the exact part count and each part's size.
What's the shortest and longest segment length I can set?
The Segment (s) field accepts 5 to 3600 seconds (one hour), in steps of 5. For cuts shorter than 5 seconds, switch to manual markers, where each cut position is a precise value down to 0.1 s.
Can it split by silence or at specific points instead?
Yes — the same tool offers On silence (cut at detected pauses) and Manual markers (cut at exact second positions) via the Split mode dropdown. See the podcast-segment guide for silence and the timestamp guide for markers.
Why is the last part not the full segment length?
Because the duration rarely divides evenly. The final part holds whatever time is left after the full-length parts. This is intentional — the tool does not pad the tail with silence.
What input formats are supported?
MP3, WAV, FLAC, M4A, OGG, and Opus are all handled by FFmpeg. The output format mirrors whichever you drop in.
Can I rejoin the parts later?
Yes. Drop the parts into the audio merger in order and it concatenates them end-to-end into one file. This is handy if you split too finely and want to recombine a subset.
Privacy first
Every JAD Audio tool runs entirely in your browser via FFmpeg (WebAssembly) and RNNoise. Your audio files never leave your device — verified by zero outbound network requests during processing.