How to mono to stereo converter — zero upload, browser-based
- Step 1Open the tool on a trusted device — Load the page on a machine you control. The first load fetches the FFmpeg WebAssembly engine; after that the processing logic is local and the audio you add never leaves the device.
- Step 2Drop the confidential mono file in — Drag your MP3, WAV, FLAC, M4A, OGG, Opus, or video file onto the dropzone. It is read into the tab's memory only — there is no upload step, no temporary server copy, and no link generated for anyone else.
- Step 3Select the Mode — Choose Duplicate (true mono → stereo) to keep the recording faithful (identical channels), or Pseudo-stereo (Haas widen ~15 ms) for perceived width. For evidentiary or legal audio, prefer Duplicate so the content is unaltered apart from channel count.
- Step 4Process locally — Run the conversion. FFmpeg builds the 2-channel output on your CPU inside the tab. No progress data, file content, or metadata is sent anywhere.
- Step 5Download to your secure location — Save the output directly to your encrypted drive or case folder. The output container matches your input so it slots into your existing chain of custody or asset-management system.
- Step 6Close the tab to clear memory — Closing or refreshing the tab discards the in-memory buffers. Because nothing was uploaded, there is no server-side copy to request deletion of — the only copies are the ones you saved.
What stays local vs. what (doesn't) leave your device
The privacy posture of this tool, grounded in how it runs.
| Data | Leaves your device? | Detail |
|---|---|---|
| Your audio file content | No | Read into browser memory; processed by FFmpeg WebAssembly locally |
| File name / metadata | No | Never transmitted; output naming happens locally |
| The FFmpeg engine code | Downloaded once | The WASM binary is fetched to run in your browser, then cached |
| Conversion result | No | Written to a local download — you choose where it goes |
Mode behaviour (privacy-neutral — both run fully local)
Both modes process in-browser; choose based on fidelity needs.
| Mode | Option | FFmpeg | Best for sensitive audio? |
|---|---|---|---|
| Duplicate | widen: false | -ac 2 | Yes — content unaltered, identical L/R |
| Pseudo-stereo | widen: true | adelay=15:all=1 on R, then merge | Only if you need width and accept the Haas delay |
Tier limits
No-upload applies on every tier; limits govern file size and length only.
| Tier | Max size | Max duration | 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-sensitive conversion scenarios, with the mode and the reason it stays compliant.
Confidential interview — keep the audio faithful
A mono field interview needs to be stereo for an editing suite, but it must remain evidentiary-faithful. Duplicate changes only the channel count, processed entirely on-device.
interview.wav (1 ch, confidential) Mode: Duplicate → -ac 2 Output: interview-stereo.wav (2 ch, L == R) Network traffic during conversion: none Server copy created: none
Legal deposition mono → stereo, offline
Air-gapped or restricted networks. Once the page and engine are cached, conversion runs with the network off — nothing could be uploaded even if it tried.
1. Load tool while online (caches FFmpeg WASM) 2. Disconnect network 3. Drop deposition.mp3 (1 ch) → Mode: Duplicate 4. Download deposition-stereo.mp3 (2 ch) Verified: zero outbound requests during step 3-4
Unreleased track stem to a stereo file
An unmastered mono stem must stay off third-party servers. Duplicate (or Pseudo-stereo for width) keeps it local; output stays lossless if you start from WAV/FLAC.
stem.flac (1 ch, unreleased) Mode: Duplicate → -ac 2 -c:a flac stem-stereo.flac (2 ch, lossless, never uploaded)
Medical dictation requiring 2-channel archive
Dictation systems sometimes archive as stereo. Convert on the clinician's machine; no PHI transits the internet.
dictation.m4a (1 ch) Mode: Duplicate → -ac 2 -c:a aac dictation-stereo.m4a (2 ch) Processing: in-browser only, no PHI upload
When you need to remove spoken PII, not change channels
If the privacy concern is the content of the speech (names, account numbers) rather than the channel layout, this tool is the wrong one — it doesn't analyse or redact speech.
Goal: bleep/redact spoken personal data → use /audio-tools/pii-redactor Goal: just make the file 2-channel privately → this tool, Mode: Duplicate
Edge cases and what actually happens
Worried the file is uploaded during processing
No uploadIt isn't. The audio is read into browser memory and processed by FFmpeg WebAssembly on your CPU. You can verify with your browser's network panel: no request carries the file body during conversion. Only the engine code is fetched (once), not your audio.
Need to work fully offline
SupportedAfter the page and FFmpeg WASM are cached on first load, conversion runs without a connection. Disconnect the network and the tool still works — which also guarantees nothing can be transmitted. A first online load is required to fetch the engine.
Want the source content provably unaltered
Use DuplicateFor legal or evidentiary audio, use Duplicate. It copies mono to both channels (-ac 2) without touching the signal, so the stereo file sums back to the exact original mono. Pseudo-stereo alters the audio with a 15 ms delay and is not appropriate where faithfulness matters.
Lossy source means re-encoding the confidential audio
Generation lossMP3/AAC/OGG/Opus inputs are decoded and re-encoded, which introduces minor generation loss — a fidelity note, not a privacy one (it still never leaves your device). For faithful archives, convert from a lossless WAV/FLAC master.
File exceeds the size or duration cap
Limit reachedFree tier: 50 MB and 30 minutes per file, enforced separately. Long depositions or interviews may need Pro (200 MB / 120 min) or Pro-media/Developer (100 GB / unlimited). Trim locally with audio-trimmer to stay within limits without uploading.
Shared or untrusted computer
CautionNo-upload protects against server exposure, but a downloaded output and any browser disk cache live on the local machine. On a shared or public computer, save to encrypted storage and clear the download afterwards. Use a trusted device for sensitive material.
Hoping to redact names/numbers in the speech
Wrong toolThis tool changes channel layout only — it doesn't analyse, transcribe, or bleep spoken content. To redact personal data in the audio itself, use pii-redactor.
Already-stereo confidential file
By designIf the file is already 2-channel, there is nothing to expand. Use channel-splitter to isolate channels or audio-merger to recombine — both also run entirely in-browser with no upload.
Frequently asked questions
Does my file really stay on my device?
Yes. The conversion runs in your browser via FFmpeg compiled to WebAssembly. Your audio is read into local memory and processed on your CPU — it is never uploaded or transmitted. Only the engine code is fetched once to make the browser able to run FFmpeg.
Can I confirm nothing is uploaded?
Yes — open your browser's developer network panel during conversion. You'll see the engine assets load once, but no request carrying your file's content. You can also disconnect the network after the page loads and the tool still converts.
Is this safe for NDA / confidential / legal recordings?
From an upload standpoint, yes: the file never leaves your device, so it satisfies policies that prohibit third-party transmission. Use Duplicate to keep the audio unaltered. Remember the saved output and browser cache still live locally, so use a trusted, encrypted machine.
Which mode keeps the recording faithful?
Duplicate. It copies the mono signal to both channels with FFmpeg -ac 2, leaving the audio identical and mono-sum safe. Pseudo-stereo applies a 15 ms Haas delay and changes the sound, so avoid it for evidentiary audio.
Can it run offline?
After the first load caches the page and the FFmpeg WebAssembly engine, yes. Disconnect the network and conversion still works — which also means nothing could possibly be uploaded.
Does it store my file or metadata on a server?
No file content or metadata is stored server-side. The processing is local. The only thing that persists locally is the output you choose to save.
What formats can I convert privately?
MP3, WAV, FLAC, M4A, OGG, Opus, and video files (audio track). All processed in-browser; the output keeps your input container.
Can it redact sensitive speech, like names or numbers?
No. This tool only changes the channel layout. To redact spoken personal data, use pii-redactor, which is purpose-built for audio PII handling.
How large a confidential file can I convert?
Free: 50 MB and 30 minutes per file. Pro: 200 MB / 120 min. Pro-media and Developer: 100 GB / unlimited duration. Size and duration are checked separately.
Is it free?
Yes, within the free tier, with no account, no watermark, and no upload.
My file is already stereo — can it still help privately?
This tool is for mono sources. For existing stereo, the sibling channel-splitter and audio-merger also run fully in-browser with no upload.
What happens to the audio when I close the tab?
The in-memory buffers are discarded. Since nothing was uploaded, there is no server-side copy — only the file you explicitly downloaded remains, wherever you saved it.
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.