How to archive mp3 as flac without uploading — browser tool
- Step 1Open the tool on a trusted machine — Load MP3 to FLAC. The FFmpeg 8.1 WebAssembly engine downloads once and caches. For maximum assurance, you can load the page, then go offline before adding files.
- Step 2Drop the private MP3 — Drag a single sensitive
.mp3onto the dropzone. It's read into the tab's memory only — there is no upload step. Free-tier files can be up to 50 MB / 30 minutes. - Step 3Choose a compression level — The Compression level control defaults to
8 - default sweet spot. Use12 - smallest, slowestto minimise archive size, or0 - fastestfor speed. The level never alters the audio. - Step 4Convert locally — FFmpeg decodes and re-encodes entirely on your CPU. No progress bar means a server queue — there is no server; the time is purely local encoding.
- Step 5Download to your secure storage — Save the
.flacstraight to your encrypted drive or secure archive location. There's no temporary cloud link to clean up afterward. - Step 6Optional: confirm nothing was sent — Re-run with DevTools' Network tab open. You'll see the cached engine and no upload of your file — independent proof the audio stayed local.
Privacy model: this tool vs. upload-based converters
Why in-browser processing changes the privacy calculus for sensitive audio.
| Concern | This tool (local) | Upload converter (cloud) |
|---|---|---|
| Audio leaves device | No | Yes (full upload) |
| Server-side copy | None | Temporary (varies) |
| Works offline | Yes (after engine loads) | No |
| Download link to expire/leak | None | Yes |
| Verifiable in DevTools | Yes — no file POST | N/A |
| Account/storage created | No | Often |
Compression level (only control) + what's preserved
The four UI levels affect size/speed only. Metadata and audio behaviour grounded in the FFmpeg pipeline.
| Setting / property | Behaviour |
|---|---|
| Level 0 - fastest | Largest file, quickest encode |
| Level 5 - balanced | Mid size and speed |
| Level 8 - default sweet spot | Small file, moderate speed (default) |
| Level 12 - smallest, slowest | Smallest file, slowest encode |
| Audio | Identical to source MP3 (lossless container) |
| Tags / art | ID3 -> Vorbis comments; art re-attached if present |
| Sample rate / channels | Preserved (no -ar/-ac) |
Audio tier limits (per file)
Per-file size and duration caps; duration is independent of size. All processing stays local at every tier.
| Tier | Max size | Max duration | Files/batch |
|---|---|---|---|
| Free | 50 MB | 30 min | 1 |
| Pro | 200 MB | 120 min | 10 |
| Pro-media | 100 GB | Unlimited | 100 |
| Developer | 100 GB | Unlimited | Unlimited |
Cookbook
Privacy-first archival workflows. The audio never leaves your machine, and the FLAC always decodes identical to the source MP3.
Embargoed interview to a local FLAC archive
A journalist archives interview MP3s to lossless FLAC without any of them touching a server.
Input: interview-source.mp3 (192 kbps, 47:00, ~64 MB) Tier: Pro (over the free 50 MB / 30 min limit) Level: 8 - default sweet spot -> interview-source.flac saved to encrypted drive. Network panel: only the cached engine; no file uploaded.
Proving it's local by going offline
The strongest privacy test: disconnect the network, then convert.
1. Load the tool page (engine downloads + caches) 2. Turn off Wi-Fi / pull the ethernet cable 3. Drop memo.mp3 and convert Result: memo.flac is produced with NO network at all. If it weren't local, it couldn't run offline.
Smallest archive for a constrained secure drive
When the encrypted archive volume is tight, level 12 minimises FLAC size. Audio is unchanged.
dictation.mp3 (mono, 64 kbps, 1:12:00) [Pro: under 200 MB / 120 min] Level: 12 - smallest, slowest -> dictation.flac at the smallest lossless size. Encodes slower, but saves space on the secure volume.
Keeping case metadata with the archive
Tags travel into the FLAC so the archived file stays self-describing without an external index.
MP3 tags -> FLAC Vorbis comments: Title = "Case 2026-0042 - Witness A" Date = 2026-03-11 Comment= "Recorded room 3" Edit or redact fields later (still local) with the [ID3 / tag editor](/audio-tools/id3-editor).
Removing audible PII before archiving
FLAC conversion preserves audio exactly — it does NOT redact speech. If a recording contains spoken PII you must remove, do that first with the dedicated tool.
Spoken PII in the recording? This tool won't remove it -- it preserves the audio verbatim. Use the [audio PII redactor](/audio-tools/pii-redactor) to handle spoken sensitive data, THEN archive the result to FLAC here.
Edge cases and what actually happens
Assuming 'no upload' still means a cloud copy somewhere
Local onlyThere is no server component to this conversion. The MP3 is decoded and FLAC-encoded by FFmpeg-WASM inside your browser tab; no copy is created off your device. You can confirm with the Network panel (no file POST) or by converting with the network disconnected after the engine loads.
FLAC conversion treated as a redaction step
Not redactionConverting to FLAC preserves the audio exactly — including any spoken names, numbers, or other PII. It is not a privacy filter for the content of the recording. To remove audible sensitive data, use the audio PII redactor first, then archive the cleaned file here.
Metadata leaking sensitive details
PreservedTags carry from MP3 to FLAC verbatim, so a Title or Comment field containing names/case numbers stays in the archive. If that's a concern, edit or strip fields with the ID3 / tag editor before or after conversion — also fully local.
Expecting FLAC to restore degraded private audio
By designLike any MP3-to-FLAC conversion, this freezes the current quality losslessly but can't recover what MP3 already discarded. For a sensitive recording you want preserved at best possible fidelity, archive it as FLAC as early as possible to avoid further MP3 re-saves.
Sensitive file over the tier size/duration cap
RejectedPer-file limits apply even though processing is local: Free 50 MB / 30 min, Pro 200 MB / 120 min, Pro-media and Developer 100 GB / unlimited. The duration cap is separate. A long embargoed recording may need Pro+ or a trim with the audio trimmer.
Browser extensions or DevTools modifying the page
CautionThe privacy guarantee assumes a clean browser. A malicious extension could in principle read page data. For maximum assurance with highly sensitive audio, use a trusted browser profile with extensions disabled, and verify offline operation.
Output FLAC larger than the private MP3
ExpectedLossless storage of decoded lossy audio is bigger than the MP3. On a constrained encrypted volume, choose level 12 to minimise size — it will still typically exceed the source MP3 but stays well under an equivalent WAV.
Cache cleared on a shared/locked-down machine
Re-downloadIf the WASM engine isn't cached (incognito, cleared cache, or a locked-down profile), it downloads again on next use. That's the engine, not your audio — your file is never part of any network request. After it loads, you can go offline.
Tab closed during a long local encode
Re-run requiredEverything runs in the tab; there's no server job to resume. Closing or refreshing mid-encode loses progress, but the original MP3 on disk is untouched — just re-run it. No partial file is left on any server because none exists.
Wanting to batch many sensitive files
Single fileThe tool handles one file at a time. For a set of confidential recordings, convert them sequentially in the same session — the engine stays cached and everything remains local throughout. No bulk upload happens at any point.
Frequently asked questions
Is my MP3 really not uploaded anywhere?
Correct. FFmpeg 8.1 runs as WebAssembly inside your browser, so the MP3 is read into the tab's memory and processed on your CPU. It is never sent to a server. You can verify this in DevTools' Network tab, where you'll see only the one-time engine download — never your file.
Can I prove the conversion is local?
Yes. Load the page so the engine caches, then disconnect from the network and run a conversion — it still completes, which is only possible if all processing is local. You can also watch the Network panel and confirm no file upload occurs.
Does converting to FLAC remove sensitive content from the recording?
No. FLAC conversion preserves the audio exactly, including any spoken PII. It's a format change, not a redaction. To remove audible sensitive data, use the audio PII redactor first, then archive the cleaned file as FLAC here.
Do the tags carry sensitive details into the FLAC?
They carry verbatim — a Title or Comment with names or case numbers will appear in the FLAC. If that's a concern, edit or strip the fields with the ID3 / tag editor (also local) before or after conversion.
Is there a cloud copy or download link that could leak?
No. The FLAC is generated on your machine and saved directly via the browser's download — there's no temporary server-side file or shareable link to expire or leak. Nothing about your audio exists off your device.
Will FLAC restore the quality of an old private recording?
No. It freezes the current quality losslessly but can't recover data the MP3 codec already discarded. For best fidelity, archive sensitive recordings to FLAC early, before further MP3 re-saves degrade them more.
Can I use this on an air-gapped or restricted machine?
Yes, once the engine has loaded. Load the page where there's network access (or pre-cache it), then operate offline. Conversions run with no network at all, which suits air-gapped or tightly restricted environments.
How large a sensitive file can I convert?
Per file: Free 50 MB / 30 min, Pro 200 MB / 120 min, Pro-media and Developer 100 GB with unlimited duration. The duration cap is separate from size. Processing stays local at every tier.
Why is the archived FLAC bigger than the MP3?
Because lossless storage of the decoded audio takes more space than the lossy MP3. Use compression level 12 to minimise it on a tight secure drive. It'll still be larger than the MP3, but smaller than an equivalent WAV.
Is the privacy guarantee affected by browser extensions?
The audio still isn't uploaded by the tool, but a clean browser is best practice for highly sensitive material. Use a trusted profile with extensions disabled and verify offline operation if your threat model warrants it.
Can I convert a batch of confidential recordings?
One file at a time. Run them sequentially in the same session; the engine stays cached and every file is processed locally. There's no bulk upload at any stage.
Are sample rate and channels kept for an accurate archive?
Yes. No resampling or channel remixing is applied, so the FLAC matches the source MP3's rate and channel layout exactly — important for a faithful archive. Change them deliberately with the sample-rate converter only if you need to.
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.