How to convert ogg to mp3 without uploading — browser tool
- Step 1Open the tool (then optionally go offline) — Load the page once. To prove nothing is uploaded, you can disconnect from Wi-Fi before converting — FFmpeg-WASM is already in the tab and needs no network.
- Step 2Drop the OGG recording in — Use the dropzone. The file is read into browser memory only. It also accepts Opus, WAV, M4A and more, so dictation apps that export OGG-Opus are covered.
- Step 3Confirm the recording's stats — The panel shows duration, sample rate and channels read locally from the file header. Confirm the duration matches the recording so you know the whole session loaded.
- Step 4Set the bitrate for speech — Choose 128k for spoken-word recordings — it is clear and compact. Use 192k+ only if the recording includes music or you need a higher-fidelity master. Bitrate is the only control.
- Step 5Run the conversion locally — Click Run OGG → MP3. Encoding happens on your CPU via FFmpeg-WASM. No bytes leave the machine; the only optional network call is a single anonymous usage counter you can disable in account settings.
- Step 6Preview, then download and securely store — Play the embedded preview to confirm the full recording is intact, then Download the MP3. Store it where your confidentiality policy requires — the tool keeps no copy.
Where confidential OGG recordings come from
Common private recording sources and the bitrate that suits them. All convert locally with no upload.
| Recording type | Typical OGG codec | Suggested bitrate |
|---|---|---|
| Journalist interview | Vorbis or OGG-Opus | 128k (speech) / 192k if ambient detail matters |
| Legal dictation | Vorbis (recorder apps) | 128k |
| Call recording | OGG-Opus (VoIP/voice apps) | 128k; often mono |
| HR / meeting capture | Vorbis | 128k–192k depending on number of speakers |
| Field / research audio | Vorbis | 192k–256k to keep environmental detail |
What stays private vs. what leaves the browser
Exactly what does and does not touch the network with this converter.
| Item | Leaves your device? |
|---|---|
| The OGG audio you load | No — read into the tab, never uploaded |
| The MP3 output | No — generated locally, downloaded directly |
| Tags / embedded metadata | No — copied locally into the MP3 |
| Anonymous 'file processed' counter | Optional — a single count, no content; can be disabled in settings |
| The conversion after page load | Works fully offline |
Cookbook
Private-recording conversions, with the privacy guarantee and the right bitrate spelled out for each.
Interview OGG to MP3 for a transcription service
A recorded interview saved as OGG. Your transcription tool only takes MP3. Convert at 128k locally so the audio never passes through a conversion server before reaching the (separately chosen) transcription step.
Source: interview-2026-06.ogg (52 min, mono, Vorbis) Option: Bitrate = 128k Output: interview-2026-06.mp3 (128 kbps mono) Privacy: decoded + encoded in-tab; no upload. Note: 52 min > 30 min Free limit -> needs Pro (120 min).
VoIP call recording (OGG-Opus) to MP3
A call-recording app exported OGG-Opus. FFmpeg decodes Opus too, so it converts to MP3 here. Calls are usually mono speech — 128k is ideal and stays small.
Source: call_2026-06-14.ogg (OGG-Opus, mono) Option: Bitrate = 128k Output: call_2026-06-14.mp3 (128 kbps mono) No bytes left the device during conversion.
Prove it's offline
The clearest privacy test: load the page, turn off your network, then convert. It still works because FFmpeg-WASM is already in the tab.
1. Open the tool, let it load 2. Disable Wi-Fi / pull the ethernet cable 3. Drop the OGG, set bitrate, Run 4. Conversion completes -> proof nothing was uploaded
Keep recorder-embedded case metadata
Some professional recorders embed date or case-reference text in the file's tags. The converter copies tags across, so the MP3 keeps that audit trail.
Source OGG tags: Title="Case 1182", Date="2026-06-14" Output MP3: same Title/Date copied to ID3v2.3 Want to scrub or change them? -> id3-editor at /audio-tools/id3-editor (also local)
Redact a section before sharing
If the recording contains PII you must remove before sharing the MP3, mute the sensitive range first, then convert. Both steps run locally.
Workflow (all in-browser, no upload): 1. pii-redactor -> mute the spoken card number range 2. ogg-to-mp3 -> encode the redacted audio at 128k Result: shareable MP3 with the PII silenced.
Edge cases and what actually happens
Does any audio reach a server?
By designNo. FFmpeg-WASM decodes and encodes inside the browser tab. The audio, the output MP3, and the metadata all stay on your device. The only optional network call is a single anonymous 'one file processed' counter (no content), which you can disable in account settings.
Recording is OGG-Opus, not Vorbis
SupportedCall-recording and voice apps frequently use OGG-Opus. FFmpeg decodes Opus as well as Vorbis, so it converts to MP3 here without issue. There is also a dedicated Opus to MP3 tool with voice-friendly defaults.
Recording is longer than 30 minutes (Free)
Tier limitInterviews and meetings often run past the Free tier's 30-minute / 50 MB cap. Either split the recording first with the audio-splitter, or upgrade to Pro (120 min / 200 MB) or Pro+Media (unlimited). Splitting also keeps each file local.
Converting does not remove spoken PII
Check workflowRe-encoding does not redact anything spoken in the audio. If the recording contains card numbers, addresses or names you must remove, mute those ranges with the pii-redactor first, then convert the redacted audio.
Metadata may carry identifying info
Review tagsThe converter copies the OGG's tags into the MP3 — including any case numbers, dates or device names a recorder embedded. If you are sharing the file, review and scrub those with the id3-editor before sending.
Quality loss from re-encoding speech
NegligibleFor speech at 128k, the second lossy generation is essentially inaudible — intelligibility is fully preserved. Going below 128k starts to introduce artifacts; 64k is only for cases where size is the overriding concern.
Browser memory limits on very long files
Hardware boundBecause everything runs in the tab, an extremely long recording is bounded by available RAM, not a server. On Free the 50 MB cap keeps you well within limits; large Pro+Media files are best handled via the desktop runner, which still processes locally.
Corrupt or partially recorded OGG
errorIf a recording app crashed mid-session the OGG may be truncated and FFmpeg can fail to demux it, showing a processing error. Sometimes the audio up to the crash point still decodes — try, and if it fails, recover the source from the recorder if possible.
Frequently asked questions
Is my recording really not uploaded?
Correct — there is no upload. FFmpeg runs as WebAssembly in your browser tab, so the OGG is decoded and the MP3 encoded entirely on your device. You can prove it by going offline after the page loads: conversion still works because no server is involved.
What's the one thing that does touch the network?
Optionally, a single anonymous counter incremented when a file is processed — just a number for your dashboard, never any audio or metadata. You can turn it off in account settings. The recording itself never leaves your machine.
Can I convert call recordings and voice memos?
Yes. Many call and voice apps export OGG-Opus, and FFmpeg decodes Opus as well as Vorbis, so they convert to MP3 here. For voice notes specifically, the dedicated Opus to MP3 tool uses voice-friendly defaults.
Does converting remove sensitive information?
No — re-encoding does not redact spoken content. If the audio contains PII you need gone, mute those segments with the pii-redactor first, then convert. Also review the file's tags, which the converter copies across, before sharing.
What bitrate is best for interviews and dictation?
128k is the sweet spot for speech — clear and compact. Use 192k or higher only if the recording has important ambient detail or music. Below 128k you start to hear artifacts; 64k is only for size-critical voice.
My interview is an hour long — can I convert it?
Not on Free, which caps at 30 minutes / 50 MB per file. Split it first with the audio-splitter (also local), or upgrade to Pro (120 minutes / 200 MB) or Pro+Media (unlimited duration). All options keep the audio on your device.
Do I need an account?
No account is needed for Free-tier files, which means nothing links your identity to the recording. An account only unlocks higher limits and the dashboard counter — both optional.
Will the date/case info my recorder embedded survive?
Yes — the converter copies tags into the MP3's ID3v2.3 metadata. That preserves an audit trail. If you would rather scrub it before sharing, use the id3-editor, which also runs locally.
Is the quality good enough for transcription?
Yes. At 128k mono, speech is fully intelligible for human and automated transcription. The second lossy pass is inaudible for voice at that bitrate. Keep the source sample rate (the tool preserves it) for best results.
Does it work in a locked-down corporate browser?
Usually yes — it is a standard web page running WebAssembly, no plugins or installs. Multi-threaded FFmpeg needs cross-origin isolation; if that is blocked it falls back to single-thread (slower but identical output). No outbound data transfer is required to convert.
Can I batch several confidential files privately?
Pro processes up to 10 files per batch and Pro+Media up to 100 — all locally, no upload. For larger volumes, the desktop runner processes on your own machine. Privacy is the same at every tier because conversion is always local.
What if the OGG is corrupt?
A truncated or damaged recording may fail to demux and show a processing error; sometimes it decodes up to the damaged point. Recover the original from the recording device if possible. Nothing was uploaded during the attempt.
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.