Interview Audio Redactor

Silence interview audio segments — protect speaker identity, remove sensitive disclosures, or mute off-the-record remarks. Runs entirely in your browser. No file upload, no server, fully private.

🔒 100% client-side — your audio never leaves your device

1. Load your recording

🎤

Click to choose or drag & drop an audio file

Supports MP3, WAV, OGG

2. Mark segments to silence

Click and drag on the waveform to select a region. Add as many regions as needed, then export.

0:00.0 / 0:00.0

✅

redacted.wav

Download

How it works

The tool decodes your audio in-browser using the Web Audio API, then zeroes out every PCM sample within the marked time ranges before encoding a new WAV file — all without ever sending your audio to a server.

1LoadYour MP3, WAV, or OGG is read by FileReader and decoded to raw PCM samples via AudioContext.decodeAudioData().

2VisualiseA waveform is drawn on a <canvas> element by sampling the decoded audio buffer's channel data at regular pixel intervals.

3MarkDrag on the waveform to select time ranges. Each range is stored as [startSec, endSec]. You can add multiple, overlapping ranges are fine.

4Redact & downloadSample values in every marked range are set to 0. The modified PCM data is encoded as a 16-bit stereo WAV and offered as a download.

The zeroing is permanent — the output WAV contains true silence in the marked regions, not just volume automation. Spectral analysis tools will confirm the samples are exactly 0.

Frequently asked questions

Will this work with my MP3 interview recording?: Yes. The browser decodes the MP3 (or WAV, OGG) into raw PCM using the Web Audio API's decodeAudioData — the same engine used by your browser's media player. The output is always WAV (PCM) because WAV is the simplest format to encode from scratch in JavaScript without an external library. If you need a smaller MP3 output, convert the downloaded WAV with any free tool such as Audacity.
Is the silencing permanent, or can someone recover the original audio?: The exported WAV file contains literal zeros in the silenced regions — there is no hidden layer, no metadata copy of the original waveform, and no reversible encoding step. The original file on your computer is never modified; only the downloaded WAV is altered. If you need to make the redaction forensically unrecoverable, delete or securely wipe your source file after exporting.
My recording is long (30+ minutes). Will the browser run out of memory?: The Web Audio API decodes the entire file into a Float32Array in RAM, which is memory-intensive. A 30-minute stereo recording at 44,100 Hz occupies about 400 MB of uncompressed PCM. Most modern desktops handle this fine. On mobile or lower-RAM devices, very long files may cause the page to slow or fail. If that happens, split the file into shorter segments using a free tool like Audacity before loading it here.
How do I select multiple non-contiguous regions?: Click "Add region (drag)" to enter region-adding mode, then drag on the waveform for each segment you want to silence. Each drag creates an independent region shown in red. You can keep adding as many regions as you like before exporting. Click the ✕ button next to any region in the list below the waveform to remove it individually, or "Clear all regions" to start over.
Does the tool handle stereo recordings?: Yes. All channels are processed — mono (1 channel), stereo (2 channels), and multi-channel files are all zeroed across every channel in the selected time ranges. The output WAV preserves the channel count of the original decoded audio (browser WebAudio typically outputs stereo; mono sources are usually expanded to stereo).