Compress Audio File Size
Reduce an audio file's size, compare the processed result with the original, and download the compressed audio. The encoder runs in your browser, so your file is not uploaded and no transcription minutes are used.
Compress one MP3, WAV, or M4A file to MP3 on this device.
The compressor detects the real container and codec, re-encodes the audio track to a chosen MP3 bitrate, and shows the finished size before you download. Your audio never travels to a WhisperWeb upload, transcription, or AI service.
Free local compression · no credits or payment required.
96 kbps stereo is the default. Bitrate is what drives file size: lower kbps means a smaller file. Mono keeps speech clear at a lower bitrate, but at the same fixed bitrate the size stays similar. A file that is already small can grow when it is re-encoded, and the result screen says so before you download.
Compression runs locally and does not use transcription credits. Re-encode spoken audio to a smaller MP3, then take it to MP3 to text when you need a transcript. Cloud speech jobs still follow plans and usage limits.
Before and after on the same 30-second sample
Download the self-made tone sample, compress it at 96 kbps stereo or 64 kbps mono, then compare the file sizes. Both results keep the full 30-second duration.
sample-source.wav
5,292,078 bytes (about 5.0 MB) · 30.0 s · 44.1 kHz stereo PCM. Uncompressed audio is large because every sample is stored, which is exactly the file you want to shrink before sending or uploading.
Download sample WAVsample-96kbps-stereo.mp3
360,848 bytes (about 352 KB) · 30.0 s · 96 kbps stereo MP3. About 93% smaller than the source, produced with the same encoder this page runs in your browser.
Download 96 kbps MP3- Source WAV
- 5,292,078 bytes · 30.0 s · PCM stereo. Both MP3 results keep this exact duration.
- 96 kbps stereo MP3
- 360,848 bytes (about 352 KB) · 360 KB target. Smaller spectral detail, still clear for voice, podcasts, and shared notes.
- 64 kbps mono MP3
- 240,579 bytes (about 235 KB). About 95% smaller than the source. Best for spoken-word audio that only needs one channel.
- Sample license
- Self-made tone sample built for this page. It is not a customer recording and not an accuracy claim. Reuse it freely to test the workflow.
How to compress an audio file
The file, the settings, the preview, and the download all stay on this page.
- 1
Choose an MP3, WAV, or M4A file
Pick one audio file up to 100 MiB. The compressor reads the real container and codec instead of trusting the file extension, then reports the detected format.
- 2
Set the target bitrate and channels
Choose 64, 96, or 128 kbps and mono or stereo. Lower bitrates and mono produce smaller files. The preflight estimate appears before encoding starts.
- 3
Compare the result, then download
Preview the compressed MP3, read the actual output size next to the original, and download it only when you are happy. If the result is larger, the page tells you before you save it.
Small files move further
Voice memos, interview recordings, podcast episodes, and lesson audio often start as large WAV or high-bitrate files. Compress a copy to fit an upload form, an email attachment, or a slow connection, and keep the full duration intact.
Illustrated use case, not a customer recording. No customer file is shown or sold on this page.
Reduce file size versus dynamic compression
Two different jobs share the word "compression". This page does the file-size one.
An audio compressor in this page means data compression: the encoder stores the same performance in fewer bytes. The waveform you hear stays the same length, and the file you export is a normal MP3 any player, phone, or upload form accepts.
A dynamics compressor is a studio effect. It turns down peaks above a threshold so the quiet and loud parts sit closer together, then often adds makeup gain. That changes how the recording moves and breathes, and it does not make the exported file smaller on its own. If you need that effect, use a mixing tool, not a file-size compressor. See the MDN audio codec guide for how lossy codecs discard detail.
Because this tool only re-encodes audio, it never removes video, never changes the running time, and never rewrites spoken words. It is a preprocessing step you run before a transcription, upload, or attachment limit, not a replacement for one.
Bitrate and quality trade-offs
MP3 bitrate is the number of bits stored per second. Higher numbers keep more detail and create a larger file.
- 64 kbps mono
- Smallest result. Good for voice memos, interviews, and notes that only need one channel. Music loses cymbals, reverb tails, and stereo width.
- 96 kbps stereo
- The balanced default. Clear speech with some stereo image, roughly one third of a 320 kbps file's size. A common choice for podcasts and e-learning voice tracks.
- 128 kbps stereo
- Closer to the original for music and mixed audio, at about twice the size of 64 kbps mono. Use it when the file is still under your destination's limit.
- What does not change
- Duration, sample order, and channel count you choose. Re-encoding an already lossy MP3 adds one more generation of loss, so keep an original copy when quality matters.
The size estimate is duration multiplied by the target bitrate. A 30-second file at 96 kbps is about 360 KB before container overhead, which is why the sample above lands near that number. VBR and metadata can move the final bytes slightly.
Compare before downloading
A smaller file is the goal, but the result is not guaranteed to be smaller.
If your source is already a low-bitrate MP3, re-encoding it at a higher target bitrate makes a larger file. If your source is a short WAV, the MP3 will usually be far smaller. The result panel always shows the original size, the generated size, and the exact percentage, so you decide with numbers instead of a promise.
Preview the compressed MP3 in the player before you keep it. If it is larger or if the quality is not good enough, change the bitrate or channel mode and run it again. The original file on your device is never modified, and nothing is uploaded while you iterate.
Keep the original, ship the smaller copy
Teams often hit an upload or email limit with a WAV or a high-bitrate recording. Compress a copy here for the destination, and archive the original in case you need to re-encode later.
Illustrated use case, not a customer recording. The downloadable samples above are the source of truth for the measured sizes.
WhisperWeb vs Happy Scribe vs cloud audio compressors
Public product pages checked 19 September 2026. Plans and limits change. Confirm the workflow you need.
| Decision | WhisperWeb | Happy Scribe | Cloud compressor apps |
|---|---|---|---|
| Where the audio goes | WhisperWebAvailable: Stays in your browserThe verified encoder runs on this device. No upload, no queue, and no server copy of the source or result. | Happy ScribeLimited: Uploaded to a serverIts public audio compressor tool describes uploading the file and waiting for a server-side job. | Cloud compressor appsLimited: Uploaded to a serverMost cloud converters upload the file, then expose a temporary download link. |
| Cost and account | WhisperWebAvailable: Free, no credits, no sign-inCompression never touches the transcription billing path, so it cannot consume transcription minutes. | Happy ScribeLimited: Free tier with limitsPublic tools commonly cap duration or size and push larger jobs toward a paid plan. | Cloud compressor appsLimited: Free tier with capsFree use often adds file-size ceilings, watermarks, or slower queues. |
| Control over the result | WhisperWebAvailable: Bitrate, channels, and a size checkChoose 64/96/128 kbps and mono or stereo, preview the MP3, and see the exact size change before downloading. | Happy ScribeLimited: Preset quality levelsPublic tools usually offer low/medium/high presets rather than an explicit kbps and channel choice. | Cloud compressor appsLimited: Preset quality levelsQuality is often a single slider, with the final bitrate hidden until after processing. |
| Speech work next | WhisperWebAvailable: One place for the transcript tooWhen you do need text, run the compressed file through the MP3 to Text converter or the workspace and review it on the page. | Happy ScribeAvailable: Transcription in the same suiteIts transcription products are the main offering; the compressor is a supporting utility. | Cloud compressor appsNot available: Compression onlyA pure converter stops at the download and sends you elsewhere for text. |
Checked 19 September 2026 from public converter copy. No outbound competitor links. No price rankings. Compression on WhisperWeb is browser-local and separate from metered speech jobs.
Choose WhisperWeb when you want a free, no-sign-in file-size reduction that keeps the audio on your device and shows the exact size change. Choose a cloud service if you need a server-side batch that runs after you close the tab. Choose a desktop editor if you also need dynamics processing or multitrack mixing. Need text instead of a smaller file? Convert the compressed MP3 with the MP3 to Text tool.
Compression here is a pre-processing step. It never triggers transcription, never reserves speech minutes, and never writes your audio into a public page or sitemap.
Limits, privacy, and cost
Accepted input is one MP3, WAV, or M4A file up to 100 MiB. The compressor probes the real container, codec, and duration before encoding, then refuses inputs without a usable audio stream. Output is always mono or stereo MP3 at 64, 96, or 128 kbps, capped at 64 MiB per result.
Encoding runs in this browser on a verified ffmpeg.wasm runtime. The file is not uploaded, not stored, and not sent to a transcription or AI service. Closing or refreshing the tab clears the in-memory copy, and downloading the MP3 is the only way to keep the result.
Compression does not use transcription credits and does not require an account. Speech tools on this site remain metered. See plans and usage limits for those jobs. We do not advertise unlimited free conversion as a billed product.
Frequently asked questions
Will every file become smaller?
No. The result depends on the source. A WAV or other uncompressed recording usually shrinks a lot, while an MP3 that is already lower than your target bitrate can grow after re-encoding. The result panel shows the original size, the generated size, and the percentage, and it warns you when the output is larger so you can keep the original.
Is this a music dynamics compressor?
No. This page performs data compression: it stores the same performance in fewer bytes at a chosen MP3 bitrate. A dynamics compressor is a mixing effect that lowers peaks and changes loudness. This tool does not apply threshold, ratio, or makeup gain, and it does not change how the performance sounds moment to moment beyond the chosen bitrate.
Does compression transcribe my audio?
No. The compressor never calls a speech model and never spends transcription minutes. It only re-encodes the audio track to MP3. When you want text, upload the compressed file to the MP3 to Text converter or the transcription workspace, which are separate metered jobs.
Can I compress a video's audio track?
This page accepts audio files: MP3, WAV, and M4A. It ignores any video stream if one is present, but the intended inputs are audio files. To pull audio out of a MOV or MP4 first, use a local file converter, then compress the extracted audio here.
Is my audio uploaded or stored?
No. The encoder runs in your browser on a verified, hash-checked ffmpeg.wasm runtime. Your audio never reaches a WhisperWeb server, and the result exists only in this tab until you download it. Refreshing the page discards the in-memory copy.
Which setting should I choose?
For spoken audio such as interviews, voice memos, and podcasts, 64 kbps mono or 96 kbps stereo keeps speech clear and files small. For music or mixed audio where you still need to shrink the file, start at 128 kbps stereo. Lower the bitrate only if the file must fit a specific upload limit.