Online Voice Recorder – Record Your Microphone in the Browser

Record audio with a live mixing channel strip — gain, EQ, compressor, and noise gate — right in your browser.

This is a full recording console, not a single record button: pick an input device, tune a ten-band equalizer, compressor and noise gate while listening to the live signal, watch peak and RMS meters that reflect audio after every processing stage, then capture the take and optionally save the whole channel-strip setup as a reusable preset. Nothing leaves your browser tab — recording, processing and storage all happen locally.

How to record your voice online

Pick a Microphone or leave System default, and load a saved setup from Settings presets if you made one earlier.
Click Start microphone and allow access; the Peak and RMS meters start moving as soon as sound comes in.
Shape the sound under Advanced channel strip (gain, 10-band EQ, compressor, noise gate), and tick Save as lossless WAV in the recording options if you want WAV instead of WebM.
Click Record, speak, then click Stop & save; the take appears in the Recordings list.
Use the download button next to the take to save the file, or Export WAV to turn a WebM take into WAV.

Requesting the microphone without breaking on a mismatch

Device constraints — echo cancellation, noise suppression, auto gain control, sample rate, and channel count — are all requested as preferred values rather than hard requirements. That distinction matters: if your hardware doesn't support the exact sample rate you asked for, the browser silently falls back to its closest match instead of throwing an error and refusing to record at all. Picking a specific input device by ID is the one constraint requested as a hard requirement, since pinning an exact device is what device selection is supposed to guarantee.

The live processing chain, in order

Audio flows through a gain stage, then a ten-band peaking equalizer at the standard ISO octave centers from 31Hz up to 16kHz, then a dynamics compressor (defaulting to a −24dB threshold, 3:1 ratio, 10ms attack and 200ms release), and finally a noise gate built on an AudioWorklet from a dedicated noise-suppression package. The level meters read the signal after all of that processing, not the raw microphone input, so what you see on the peak and RMS meters is genuinely what the gate and compressor are letting through, floored at −90dBFS below which the meter just reads silence.

The noise gate can't be retuned live the way the EQ and compressor can — its open and close thresholds and hold time are fixed at the moment the underlying node is created, so changing gate settings rebuilds that stage of the chain fresh rather than adjusting a running parameter.

What actually gets recorded

Recording captures through MediaRecorder using WebM with Opus compression when the browser supports it, falling back to that browser's own default recording format otherwise — an on-screen notice appears specifically when Opus-in-WebM isn't available, so you know which path you're on. There's also an optional immediate WAV export at save time: the just-recorded WebM/Opus take is decoded back to raw samples and re-encoded as uncompressed WAV. That step doesn't add back any detail the Opus compression already discarded — it only repackages the already-lossy audio into an uncompressed container, which is useful for compatibility with software that can't read WebM, not for recovering quality.

Presets and your recording history

The entire channel-strip configuration — gain, all ten EQ bands, compressor settings, gate settings — plus the native device toggles it was captured alongside can be saved as a named preset and reloaded later to restore the whole setup in one action, not just the numeric values alone. Recordings and presets are kept in two separate local project histories, each with its own rename, duplicate and delete support, so a multi-MB recording blob and a handful of preset numbers aren't forced to share one storage shape.

When it can't get the microphone at all

If the browser denies microphone access — permission refused, no microphone present, or the device is already in use by another application — the tool surfaces the browser's own error message directly in a visible warning strip rather than translating it into a generic failure. That's deliberate: the exact wording differs by browser and by failure cause, and passing it through unchanged is more useful for figuring out what actually went wrong than a one-size-fits-all message would be.

Voice takes are free up to 30 minutes each, and Pro records for as long as you like.

What people record with a voice recorder

Voice memos and quick spoken notes to yourself.
Voiceovers for slides or videos, saved as WAV for a video editor.
Podcast drafts and interview practice before the real session.
Singing or instrument practice, played back straight from the Recordings list.
Pronunciation practice in a language you are learning.

Tips for a cleaner voice recording

Keep the microphone at a steady distance, roughly a hand's width to a forearm's length from your mouth, and speak across it rather than straight into it to soften pops.
Record in a small room with soft furnishings, and switch off fans or anything else that hums.
Watch the Peak meter: speech peaking somewhere around −12 to −6 dBFS leaves room so loud words don't clip at 0.
If the RMS reading stays very low while you talk, raise the Gain slider (it runs from −24 to +24 dB) instead of raising your voice.
Keep the noise gate on for speech in a noisy room, but consider turning it off for music with quiet passages, where it can cut soft notes.
Wear headphones if you enable Monitor / loopback playback, otherwise your speakers feed straight back into the microphone.
Once a setup sounds right, save it as a named preset so the next session starts from the same settings.

Questions people ask about Voice/Sound Recorder

Why does my recording sound processed even though I haven't touched any of the EQ or compressor sliders?: The default channel-strip settings aren't neutral — the compressor starts active at a −24dB threshold with a 3:1 ratio, and the noise gate is enabled by default too. If you want a completely unprocessed capture, you'll need to flatten the compressor and disable the gate before recording, rather than relying on default settings to be a pass-through.
Does exporting to WAV improve the quality of my recording?: No. The initial capture through MediaRecorder is WebM/Opus, which is lossy, and the WAV export step only decodes that already-compressed audio back to raw samples and repackages it uncompressed — it can't restore detail the Opus encoding already discarded. WAV export is useful for compatibility with software that doesn't handle WebM, not for recovering audio quality.
Why did the tool ask for 48000Hz but my recording came back at a different rate?: Sample rate is requested as a preference, not a strict requirement, specifically so a hardware mismatch doesn't stop recording from working at all. If your microphone or driver doesn't support the requested rate, the browser automatically substitutes its closest available match instead of failing outright.
Can I get my exact EQ and compressor setup back for a future recording session?: Yes — save your current settings as a named preset, which stores every EQ band, the compressor and gate configuration, and the native device toggles together as one unit. Loading that preset later restores the entire setup in a single action rather than requiring you to re-enter each value by hand.
Can I save my recording as an MP3?: Not directly. Takes are saved as WebM with Opus audio, or your browser's default recording format, and can be exported as WAV. For an MP3, export the take as WAV and convert it with the Audio Converter, which also works inside your browser.
How long can I record?: A single take runs for up to 30 minutes on the free tier, after which the recorder stops and saves what it captured; Pro removes that cap. Saved takes also use your browser's storage, and the Recordings panel shows the running total.
Can I record on my iPhone or Android phone?: Recording needs microphone access through getUserMedia plus the MediaRecorder API in the phone's browser. If that browser can't produce WebM/Opus, a notice explains the take will use its default recording format instead; support hasn't been confirmed on specific phones.
Is my recording uploaded anywhere?: No. The audio is processed by Web Audio nodes inside the tab, and finished takes are kept in this browser's IndexedDB so they are still listed when you come back. Clearing the site's data deletes them, so download anything you want to keep.
Can I trim the recording afterwards?: The recorder has no trim control of its own. Export the take as WAV, then open it in the Audio Cutter, which trims on a waveform in your browser and saves the shortened clip in a format of your choice.

Related tools on BrakeTools

Audio Converter — Convert between WAV, FLAC, AAC, OGG, and MP3 with title/artist/cover-art metadata carried over.
Audio Cutter — Drag in an audio file, select a range on the waveform or by typed start/end seconds, then remove or keep just that portion.
Audio Volume Changer — Make audio louder or quieter, normalize peaks, protect against clipping, add edge fades, and export in five formats without uploading it.
Screen Recorder — Record a browser tab, app window, or entire screen with optional system audio and microphone narration, entirely in your browser.
Webcam Video Recorder — Record webcam video or take full-resolution photos with quality, camera, microphone, echo-reduction, and mirror controls in your browser.

Content reviewed 2026-09-27.