Compress Audio Online: Smaller Files, Same Listening Experience

Reduce MP3, WAV, OGG or M4A file sizes by 50-90% with full control over bitrate, sample rate and channels. ffmpeg.wasm runs entirely in your browser, audio never leaves your device and you keep every kilobyte you save.

100% Private

Custom Bitrate

Size Preview

No Watermark

Setup

How to Compress Audio Files

Three steps, all client-side via ffmpeg.wasm.

Drop your audio

Drag-and-drop one or many MP3, WAV, OGG, M4A, FLAC files. The tool reads them locally, no server upload, no waiting on bandwidth.

Pick a preset or fine-tune

Choose Podcast (96 kbps mono), Standard (128 kbps stereo), High (192 kbps stereo) or Custom (set bitrate 32-320 kbps, sample rate 8/16/22/44.1/48 kHz, mono/stereo).

Compress and download

Hit compress, ffmpeg.wasm crunches the audio in your browser. See exact size delta (Before 12.4 MB → After 2.8 MB, 77% smaller). Download one file or batch as a ZIP.

Compression Without the Cloud

Every option a desktop encoder gives you, no upload required.

Bitrate Control 32-320 kbps

Pick exact target bitrate. 32-64 kbps for voice memos (intelligible but compressed). 96-128 kbps for podcasts (transparent for voice). 192-320 kbps for music (audiophile-grade).

Mono / Stereo Conversion

Podcasts and lectures rarely need stereo, switching to mono cuts file size by 50% with zero quality loss for voice content. Music keeps stereo for spatial separation.

Sample Rate Downsampling

Default 44.1 kHz is overkill for voice. Drop to 22.05 kHz for memos (50% smaller), 16 kHz for telephone-grade speech (75% smaller). Music stays at 44.1 or 48 kHz.

Live Size Preview

Before pressing compress, the tool estimates the output file size based on duration × bitrate × channels. Tweak settings, see the new estimate, find the right balance without trial-and-error.

Batch Processing

Drop 50 podcast episodes at once. Each gets the same settings, output is a single ZIP with every compressed file inside. Great for archiving a back-catalog.

100% Local via ffmpeg.wasm

Industry-standard ffmpeg compiled to WebAssembly runs in your browser tab. Audio data, original or compressed, never touches our servers. Confirm with the Network tab in DevTools.

Audio Compression: The Theory

Bitrate, codecs and what 'lossy' actually means for your ears.

Bitrate is the dial that matters most

Bitrate (kbps = kilobits per second) directly controls quality vs file size. Halve the bitrate, halve the file size, more aggressive compression artifacts. For voice content, 64-96 kbps mono is transparent to most listeners. For music, 192-256 kbps stereo is the sweet spot, 320 kbps is overkill and only audiophiles can hear the difference in blind tests. Going below 64 kbps for music produces audible 'underwater' artifacts.

Podcasters: 96 kbps mono saves 70% disk space vs 192 kbps stereo with no audible loss for voice.

Sample rate: why 44.1 kHz isn't always needed

Sample rate caps the highest frequency audio can represent (half the sample rate, per Nyquist). 44.1 kHz captures up to 22.05 kHz, the limit of human hearing, fine for music. 22.05 kHz captures up to 11 kHz, plenty for voice (which lives mostly in 100 Hz to 8 kHz). 16 kHz is telephone quality. Cutting sample rate halves file size while preserving voice content quality.

Voice memo at 16 kHz / 64 kbps mono = around 0.5 MB/min. Same content at 44.1 kHz / 192 kbps stereo = around 1.4 MB/min, 3x bigger for zero perceived quality gain.

Mono vs stereo, when stereo is wasted

Stereo doubles the file size by storing two independent channels. Stereo matters for music, ambisonic field recording, and binaural recording. Stereo is wasted for solo voice (podcasts, lectures, voice memos), single-mic recordings, and old mono content (radio broadcasts pre-1965). Converting wasted stereo to mono cuts file size in half with zero audible difference, the channels were already identical.

Run 'is the L channel different from R?' If no, it's already mono in stereo packaging. Convert to true mono and reclaim 50% disk.

Lossless vs lossy, when each makes sense

Lossless formats (WAV, FLAC, ALAC) store exact bit-for-bit audio data, big files, perfect for editing/archival. Lossy formats (MP3, AAC, OPUS, OGG Vorbis) drop psychoacoustically inaudible data, smaller files, perfect for distribution. AAC at 128 kbps usually sounds better than MP3 at 128 kbps. OPUS is the modern champion for voice (best quality at lowest bitrate). We default to MP3 because of universal device support, switch to OPUS or AAC if your audience has modern devices.

Archive in WAV/FLAC. Distribute in MP3 or AAC. Never re-encode a lossy file from a lossy source, always go back to the lossless master if possible.

Real-world use cases

Podcast distribution prep

Master at 256 kbps stereo, compress distribution version to 96 kbps mono. RSS feed bandwidth drops 70%, episode quality unchanged for listeners.

Email-friendly audio

Email providers cap attachments at 20-25 MB. A 30-minute voice memo at WAV is around 300 MB, compressed to 96 kbps MP3 it's 22 MB. Fits Gmail, Outlook, ProtonMail.

Voice memo archive

Years of voice memos at full quality eat gigabytes. Bulk compress to 64 kbps mono, archive shrinks 10x with no perceived loss for old voice notes.

Web playback optimization

Embedded audio on your blog or course page? Heavy MP3s slow page load. Compress to 96-128 kbps, listeners get same experience, your CDN bill drops.

Audiobook for travel

Audiobook MP3s often run 192 kbps stereo, around 100 MB per hour. Recompress to 64 kbps mono for travel (no audible loss for narration) and a 20-hour book fits on a 1.5 GB SD card.

Discord / Slack voice clip

Discord caps file uploads at 8 MB (free tier). Voice clip over the limit? Compress to 64-96 kbps and it fits, audio quality is fine for casual chat.

Compress Audio, Frequently Asked Questions

Quick answers about the tool

How much smaller can my audio get?

Depends on source and target. A 320 kbps stereo MP3 compressed to 96 kbps mono shrinks 85%. A WAV recorded at 44.1 kHz / 16-bit converted to MP3 at 128 kbps shrinks 90%. Already-compressed MP3s (128 kbps) re-encoded at the same bitrate only shrink 0-5% and lose quality. Always compress from the highest-quality source you have.

Will the audio sound worse?

Only if you over-compress. At sensible bitrates (96+ kbps voice, 192+ kbps music) the difference is inaudible in blind tests. Push too low (32-48 kbps for music) and you'll hear 'underwater' or 'metallic' artifacts.

What format should I export?

MP3 for maximum compatibility (every device on Earth plays MP3). OPUS for best quality at low bitrate (modern browsers, podcast apps). AAC for Apple ecosystem and best MP3 alternative. OGG Vorbis for open-source projects. When in doubt, use MP3.

Can I batch compress lots of files?

Yes, drop 10 or 100 files at once, the tool applies the same settings to each and gives you a single ZIP with all compressed outputs. Useful for podcasters compressing a back-catalog or archivers shrinking field recordings.

Is the file uploaded to compress?

No. ffmpeg.wasm is a WebAssembly build of the industry-standard ffmpeg encoder that runs inside your browser tab. Your audio never reaches our servers, confirm in your browser's Network DevTools panel during compression.

How long does compression take?

Roughly 0.5-3x real-time on a modern laptop, a 30-minute podcast compresses in 15-90 seconds. Mobile devices are slower, expect 2-5x real-time on a phone. Batch jobs queue serially to avoid running out of RAM.

What's the difference between CBR and VBR?

CBR (Constant Bitrate) keeps the same kbps the whole file, predictable size, slightly worse quality at any given bitrate. VBR (Variable Bitrate) uses more bits for complex passages, fewer for silence, around 10-20% smaller for same perceived quality. We default to VBR, switch to CBR if your target platform requires it (some old podcast hosts).

Can I keep ID3 tags (title, artist, album)?

Yes, by default we copy ID3 metadata from source to compressed output. Title, artist, album, year, genre and cover art all carry over. If you want to edit tags first, use our MP3 Tag Editor tool, then compress.

What if my audio gets distorted after compressing?

Two common causes: (1) bitrate too low for the content, raise to 128 kbps or higher. (2) Source was already compressed multiple times (generation loss), find the lossless original. Never compress an already-compressed file at the same bitrate.

Does it work on iPhone?

Yes, on iOS Safari 16+. ffmpeg.wasm runs on mobile but is significantly slower than desktop. Compress small files (under 50 MB) on mobile, batch jobs are better on a laptop.