How to mp3 to wav converter — zero upload, browser-based
- Step 1Load the converter (then you can go offline) — Open the mp3-to-wav tool. Once the page and the FFmpeg WASM engine have loaded, the conversion itself needs no network — you can disconnect and it still works.
- Step 2Drop the confidential MP3 — Drag a single
.mp3onto the tool. It is read into memory in the tab; nothing is transmitted. The tool handles one file at a time. - Step 3Pick the sample rate — Choose 44.1 kHz, 48 kHz, 96 kHz, or 16 kHz. For a transcription pipeline, 16 kHz is common; for general playback/archival, keep the source rate (usually 44.1 kHz).
- Step 4Pick stereo or mono — Default is Stereo. For a single-speaker recording headed to transcription, Mono halves the size and is what most speech engines expect.
- Step 5Convert locally — Run it. FFmpeg WASM decodes the MP3 to
pcm_s16leWAV on your CPU. Processing time scales with file length, but the file never leaves your machine. - Step 6Save the WAV — Download the result. Because there was no server round-trip, there is no remote copy to delete afterward — the only copies are the ones on your device.
Where your file goes (privacy model)
Comparison of a typical upload-based converter vs this in-browser tool. Verified: the engine runs as FFmpeg WASM in the tab; only an anonymous processed-counter is recorded for signed-in users.
| Step | Typical upload site | This tool |
|---|---|---|
| File transmission | Uploaded to their server | Never leaves the browser tab |
| Where it's decoded | On their server | On your CPU via FFmpeg WASM |
| Temp server copy | Yes (retention varies) | None |
| Result delivery | Download link / email | Direct in-page download |
| What's logged | Varies — often the file | Anonymous 'file processed' count only; opt-out available |
| Works offline | No | Yes, after the page loads |
Real conversion options
The only controls the tool exposes. Output is always 16-bit PCM WAV — no bit-depth control exists.
| Control | Choices | Default | FFmpeg arg |
|---|---|---|---|
| Sample rate | 44.1 / 48 / 96 / 16 kHz | 44.1 kHz | -ar <hz> |
| Channels | Stereo / Mono | Stereo | -ac <1|2> |
| Bit depth | (fixed) 16-bit | 16-bit | encoder pcm_s16le |
| Format | (fixed) WAV | WAV | encoder pcm_s16le |
Tier limits (checked locally, before processing)
Size and duration caps apply per file. Since there's no upload, these are enforced in the browser before the WASM engine runs.
| Tier | Max size | Max duration | Files |
|---|---|---|---|
| Free | 50 MB | 30 min | 1 |
| Pro | 200 MB | 120 min | 10 |
| Pro-media | 100 GB | Unlimited | 100 |
| Developer | 100 GB | Unlimited | Unlimited |
Cookbook
Conversions framed around confidentiality. In every case the audio stays on your device — these examples differ only in settings.
Confidential interview MP3 → WAV for in-house transcription
An HR or research interview was recorded as MP3. Policy forbids uploading it to third-party sites. Convert to 16 kHz mono WAV locally, then feed it to your own transcription tool.
Input: interview.mp3 (96 kbps, 44.1 kHz, stereo, 48:00)
Options: Sample rate = 16 kHz · Channels = Mono
Output: interview.wav (16-bit PCM, 16 kHz, mono)
Stays on the analyst's laptop; nothing uploaded.
Note: 48 min is within Free (30 min)? No -> 48 min needs Pro (120 min) or higher.NDA-covered demo track decoded without leaving the device
A musician sent an unreleased MP3 under NDA. You need a WAV to load into a tool, but cannot risk an upload. Convert at the source rate, stereo, in the tab.
Input: demo-nda.mp3 (320 kbps, 44.1 kHz, stereo, 3:10)
Options: Sample rate = 44.1 kHz · Channels = Stereo
Output: demo-nda.wav (16-bit PCM, 44.1 kHz, stereo)
No server copy exists to subpoena, leak, or forget to delete.Legal recording converted offline
A deposition or call recording in MP3 must be preserved as WAV for an evidence archive, with a strict no-cloud rule. Load the page, disconnect from the network, then convert.
Workflow: 1. Open /audio-tools/mp3-to-wav, let it load. 2. Disconnect Wi-Fi / Ethernet. 3. Drop recording.mp3, set 44.1 kHz / Stereo, convert. 4. Reconnect later if you wish — the WAV was made fully offline.
Medical voice memo to mono WAV
A clinician's dictation MP3 needs to become a mono WAV for an internal pipeline. Single speaker, so mono is appropriate and smaller.
Input: dictation.mp3 (128 kbps, 44.1 kHz, stereo, 4:30)
Options: Sample rate = 16 kHz · Channels = Mono
Output: dictation.wav (16-bit PCM, 16 kHz, mono)
Handled entirely on the workstation.Verify nothing left the tab
For a privacy review, confirm there's no network call during conversion. Open the browser's network tab before processing and watch.
DevTools -> Network panel -> clear -> run conversion. Expected: no XHR/fetch carrying the audio payload to a remote host. The FFmpeg WASM binary may already be cached; the file bytes are not sent.
Edge cases and what actually happens
Assuming files are deleted from a server
No server copyThere is nothing to delete — the file is never uploaded. The only copies are the source on your device and the WAV you download. That's the whole point of the in-browser model.
Network drops mid-conversion
UnaffectedExpected: once the page and WASM engine are loaded, conversion is local. A dropped connection doesn't interrupt or corrupt the conversion. You can deliberately go offline first.
Confidential file exceeds the tier limit
rejectedSize/duration caps are enforced in the browser before processing. A 48-minute interview exceeds Free's 30-minute cap and needs Pro (120 min) or Pro-media/Developer (unlimited). The file still never uploads — it's simply not processed until you're on a sufficient tier.
Expecting 24-bit for an evidence archive
16-bit onlyOutput is 16-bit PCM. If your archival standard mandates 24-bit, this tool can't meet it — the source MP3 has no >16-bit content anyway, but the format requirement is what it is.
WAV is far larger than the MP3 — disk space
By designUncompressed PCM is ~10.1 MB/min at 16-bit/44.1 kHz/stereo. A long interview becomes a large WAV. Ensure local disk headroom; nothing is offloaded to a server.
Browser memory limit on a multi-hour recording
errorBecause everything is in-tab, very long files can exhaust browser memory during WASM processing. Split the recording first with audio-trimmer or process in sections.
WASM not supported / blocked
unsupportedThe tool needs WebAssembly. All current browsers support it, but a locked-down enterprise policy could block it. Use an up-to-date Chromium/Firefox/Safari build.
Quality not recovered by going to WAV
ExpectedWAV stores the already-lossy MP3 audio uncompressed. It does not restore detail. For confidential audio this is usually fine — you want a faithful, portable, uncompressed copy, not a remaster.
Multiple confidential files at once
single fileThe tool processes one file at a time. Convert each separately. For combining, use audio-merger after converting each to WAV — still entirely in-browser.
Anonymous counter and privacy
SupportedFor signed-in users a single 'file processed' counter is recorded for dashboard stats — never the audio, filename, or content. You can opt out in account settings.
Frequently asked questions
Is my MP3 really not uploaded?
Correct. The conversion runs as FFmpeg 8.1 WebAssembly inside your browser. The file bytes are decoded on your CPU and the WAV is generated locally. You can confirm in the browser's network panel that the audio payload isn't sent anywhere.
Does it work offline?
Yes — after the page and the WASM engine have loaded, the conversion needs no network. You can disconnect and still convert, which is useful for strict no-cloud policies.
What gets logged?
For signed-in users, a single anonymous 'file processed' counter for dashboard stats — never the audio, filename, or content. You can opt out in account settings. Signed-out, nothing about the file is recorded.
Is there a server copy I need to delete afterward?
No. Since nothing is uploaded, there's no remote copy. The only copies are the source MP3 and the WAV you save, both on your device.
What format is the output?
16-bit PCM WAV (pcm_s16le) at your chosen sample rate and channel count. There is no bit-depth selector — it's always 16-bit.
Can I convert a confidential file larger than 50 MB?
Free tier caps at 50 MB / 30 minutes per file. Pro raises it to 200 MB / 120 minutes; Pro-media and Developer to 100 GB with unlimited duration. The cap is checked locally before processing — the file still never uploads.
Is this safe for legal or medical recordings?
The conversion never transmits the file, which removes the upload-exposure surface. Your own retention, access-control, and chain-of-custody obligations still apply to the files on your device — the tool just doesn't add a cloud copy.
Why would a privacy-conscious user want WAV instead of keeping the MP3?
WAV is uncompressed and decode-stable, which is what transcription tools, archives, and audio analysis pipelines typically expect. Converting locally lets you feed those pipelines without uploading the source.
Can I get a smaller file for a long recording?
WAV is uncompressed, so it will be large. If size matters more than uncompressed fidelity, convert back to a compressed format with wav-to-mp3 or to lossless-compressed FLAC via wav-to-flac — both also run in-browser.
Does converting improve the audio quality of a low-bitrate recording?
No. The MP3's lost detail can't be recovered. WAV just stores the existing audio uncompressed. For a noisy voice recording, a denoiser like the ai-noise-reducer addresses noise, not lost fidelity.
Can I batch-convert privately?
The tool is single-file, so batch isn't available here. You can convert files one after another — each stays in-browser. There's no queue sent to a server.
What if the page can't load the WASM engine?
WebAssembly must be enabled. Every modern browser supports it by default; a restrictive enterprise policy could block it. Use a current Chromium, Firefox, or Safari build.
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.