How to change audio sample rate without uploading — free tool
- Step 1Open the converter (optionally go offline) — Open the sample-rate-converter tool. Once the page and its WASM engine have loaded, you can even disconnect from the network — processing is fully local.
- Step 2Drop the confidential file — Drag in the recording (
.wav,.mp3,.m4a,.flac,.ogg,.opus,.aiff). It is read into browser memory only — there is no network request carrying the file. - Step 3Confirm it stayed local — Check your network tab if you like: no upload of the audio occurs. The file card shows the detected sample rate, read in-browser.
- Step 4Run the conversion — Process the file. It is resampled to 44.1 kHz and written as a 16-bit PCM WAV, entirely within the page. There is no rate or format selector.
- Step 5Download the result — Save the 44.1 kHz WAV to your own machine. The original and the output both live only on your device.
- Step 6Need a different target rate, still private? — For 16 kHz / 8 kHz / 48 kHz targets use JAD's programmatic API (same FFmpeg engine, full 8–192 kHz ladder), or video-to-wav for video sources — see the no-upload downsample guide.
Upload-based converters vs JAD's in-browser model
Why 'no upload' matters for confidential audio.
| Concern | Typical online converter | JAD in-browser |
|---|---|---|
| Where the file goes | Uploaded to their server | Stays in your browser memory |
| Third-party copy | Yes — subject to their retention | None — no copy is transmitted |
| Works offline | No (needs the server) | Yes, after the page loads |
| Account / sign-in | Often required | Not required |
| Compliance risk | You exported regulated data | Data never left your control |
What the private in-browser tool outputs
Verified against the shipped client and processor. No options UI is rendered for this tool.
| Property | Value | Notes |
|---|---|---|
| Output sample rate | 44,100 Hz (fixed) | -ar 44100; no on-page rate picker |
| Output format | WAV, 16-bit PCM (pcm_s16le) | No on-page format picker |
| Resampler | FFmpeg 8.1 swresample | Same engine as desktop FFmpeg |
| Channels | Preserved | Mono stays mono, stereo stays stereo |
| Network use for the file | None | Processed entirely in-browser via WASM |
Tier limits for this tool
From lib/tier-limits.ts (audio family). Everything still runs locally regardless of tier.
| Tier | Max file size | Max duration | Files per run |
|---|---|---|---|
| Free | 50 MB | 30 min | 1 |
| Pro | 200 MB | 120 min | 10 |
| Pro + Media | 100 GB | Unlimited | 100 |
| Developer | 100 GB | Unlimited | Unlimited |
Cookbook
Confidential-audio scenarios where uploading is not an option. The web tool outputs 44.1 kHz WAV, all local.
Privileged legal interview to CD rate
A recorded witness interview needs to go into a 44.1 kHz workflow. Uploading privileged audio to a random converter is a non-starter; resample it locally instead.
Input: interview_48k.wav 48,000 Hz, 2ch (privileged) Output: interview_48k.wav 44,100 Hz, 2ch, 16-bit PCM Network: no file upload — processed in the browser.
Offline conversion on an air-gapped laptop
Load the page once, disconnect from the network, then process. Because FFmpeg runs as WASM in the tab, the conversion completes with no connectivity.
1. Load tool page (online). 2. Disconnect network. 3. Drop file → convert → download. Result: 44,100 Hz WAV, produced fully offline.
Support call recording with PII
A support call full of names, card hints, and addresses needs a rate change before archiving. Keep it on the device.
Input: call_2026-06-10.flac 48,000 Hz Output: call_2026-06-10.wav 44,100 Hz, 16-bit PCM For masking spoken PII, see pii-redactor (separate tool).
Medical dictation standardised locally
Clinic dictation captured at 48 kHz needs to match a 44.1 kHz transcription tool. Resample without sending PHI anywhere.
dictation_48k.m4a 48,000 Hz → dictation_48k.wav 44,100 Hz, 16-bit PCM No upload; nothing retained off-device.
Different target rate, same privacy guarantee
Need 16 kHz for an in-house transcription model? The web page only does 44.1 kHz, but JAD's API uses the same local-first FFmpeg engine for other rates.
Web page: 44.1 kHz only. Need 16 kHz privately? Use the API (sampleRate 16000) — same FFmpeg engine, full 8-192 kHz ladder.
Edge cases and what actually happens
Assuming there must be a server step
No uploadThere is none for the file. The audio is read into browser memory and processed by WASM FFmpeg in the page; no network request carries the file. You can verify in your browser's network panel.
Wanting to work fully offline
SupportedAfter the page and its WASM engine have loaded, you can disconnect from the network and still convert. Nothing about the resample requires connectivity.
Expecting a rate or format option for sensitivity reasons
Not presentThe tool outputs a fixed 44.1 kHz 16-bit WAV with no picker. For other private targets, the programmatic API (same FFmpeg engine) covers the full 8–192 kHz ladder.
File larger than the tier cap
RejectedFree is 50 MB / 30 min; Pro is 200 MB / 120 min. The limit is enforced locally before processing — it does not imply any upload. Split long confidential recordings with audio-splitter, which is also fully in-browser.
Need to redact spoken PII, not just resample
Different toolResampling does not remove sensitive content. To mute/redact spoken PII ranges, use pii-redactor — also browser-local — then resample.
Want a compressed, still-private output
Two-step (local)Output here is WAV. To compress while staying local, follow with wav-to-mp3 or wav-to-flac — both run in-browser, so the file still never uploads.
Safari can't decode FLAC/Opus for the preview
PreservedSafari's Web Audio decoder may not render the on-page waveform/metadata for some FLAC/Opus files, but FFmpeg decodes and resamples them independently — locally and without upload.
Source already 44.1 kHz
No-opRe-running a 44.1 kHz file produces another 44.1 kHz WAV with no benefit. Check the file card and skip files already at the target.
Worried about browser memory on a huge file
Tier-boundedBecause everything is in-memory, very large files are bounded by tier caps (e.g. Pro 200 MB). If a confidential file is too large, split it locally with audio-splitter first.
Frequently asked questions
Is my audio really not uploaded?
Correct. The tool runs FFmpeg 8.1 compiled to WebAssembly inside your browser tab. The file is read into local memory and processed there; no network request transmits the audio. You can confirm this in your browser's network panel.
Can I use it completely offline?
Yes, once the page and its WASM engine have loaded. You can disconnect from the network and still drop a file, resample, and download — the conversion needs no server.
Why does no-upload matter for my recordings?
Privileged legal interviews, medical dictation/PHI, HR investigation audio, and support calls with PII should not be exported to a third party just to change a sample rate. Uploading creates a copy governed by someone else's retention policy — a real compliance exposure. Local processing avoids that entirely.
What does it output?
A 44.1 kHz, 16-bit PCM WAV. There is no rate or format dropdown for this tool — 44.1 kHz WAV is the fixed result, which suits the common 'standardise to CD rate' need.
Can I get 16 kHz or 48 kHz and stay private?
Yes — through JAD's programmatic API, which uses the same local-first FFmpeg engine and supports the full 8–192 kHz ladder. The in-browser page itself only outputs 44.1 kHz.
Is the resampling lower quality because it's in the browser?
No. It is the same FFmpeg swresample library as desktop FFmpeg, compiled to WebAssembly. The result is identical to running ffmpeg -i in -ar 44100 out locally.
Does it keep my channel count and metadata?
Channel count is preserved, and tags/cover art are carried where the container supports them (-map_metadata 0). For richer tagging afterward use id3-editor — also browser-local.
How do I redact sensitive speech, not just change the rate?
Resampling does not remove content. Use pii-redactor to mute specified time ranges (also fully in-browser), then resample if needed.
What if my confidential file is very large?
Limits are enforced locally: Free 50 MB / 30 min, Pro 200 MB / 120 min, Pro + Media and Developer 100 GB. Nothing is uploaded regardless. Split oversized files with audio-splitter first.
Can I compress the output and still avoid uploading?
Yes. Resample here to 44.1 kHz WAV, then run wav-to-mp3 or wav-to-flac — both are in-browser, so the audio still never leaves your device.
Does anything get logged or retained on a server?
The audio file is never transmitted, so there is nothing to retain. There is no account requirement for the conversion itself, and closing the tab clears the file from memory.
Which related tools are also no-upload?
All JAD audio tools run in-browser, including pii-redactor, audio-splitter, wav-to-mp3, wav-to-flac, and video-to-wav. See also the downsample guide and the 48→44.1 guide.
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.