Voice Recorder

How do you record, trim, and download a voice memo?

Press Start recording, allow microphone access, speak, and press Stop. The finished take offers Download original for the complete compressed recording and Download WAV for the selected waveform range. Use Choose a microphone first if you need a particular input instead of the default. There is no audio upload step, transcription field, or automatic silence remover: you choose the recording boundaries yourself.

For a worked example, record 12 seconds of speech, press Pause for a 5-second break, then press Resume and record another 18 seconds. The elapsed counter reaches approximately 00:30 rather than 00:35 because paused time is excluded. After Stop, suppose the decoded take is exactly 30 seconds long. Moving the selection start to 2 seconds and its end to 28 seconds retains 26 seconds. Play selection checks that range; Download WAV saves it. Download original is disabled until Reset trim restores the full take.

If that decoded selection has one channel at 44,100 samples per second, its 16-bit WAV contains 26 × 44,100 = 1,146,600 sample frames. Each mono frame needs 2 bytes, giving 2,293,200 bytes of sample data plus the 44-byte WAV header: 2,293,244 bytes total. Two channels at the same rate and duration produce 4,586,444 bytes. These are conditional examples, not a promise about your microphone's channel count or the decoded sample rate.

How do you prepare a useful voice memo or interview?

Start with a short test rather than your full conversation. Choose a microphone, speak at the distance you intend to use, and record a few sentences. Stop and listen to the result before committing to a longer take. For a personal memo, state the subject at the beginning so the saved file is easy to identify later. For an interview, obtain the participants' agreement before recording and position the microphone where everyone can be heard clearly.

Pause recording between sections when you want one continuous file without lengthy breaks. The elapsed timer excludes paused time. Resume adds more audio to the same take; Stop finishes it and turns off microphone capture. Keep this tab open throughout the session. A supported screen wake lock helps keep the screen awake, but operating system restrictions, background tabs, and device sleep can still interrupt capture. Save important material in shorter takes and download each one promptly.

How can you choose the right microphone and input level?

Press Choose a microphone to request permission and reveal the available inputs. The permission check releases its microphone stream immediately. Select a built-in microphone, headset, or connected interface before starting the take. If an expected input is missing, check the cable, device settings, and browser permission. The default microphone follows your system's current default; selecting a named device requests that particular input.

Watch the level indicator while speaking. The labels follow absolute sample peaks: low is below 0.02, ok is at least 0.02 but below 0.97, and clipping starts at 0.97. These thresholds describe the meter, not a guarantee of intelligible speech or undistorted audio. Move closer if the level remains low, or reduce your interface gain and increase your distance if it clips. The meter observes the captured signal; it does not repair distortion already introduced at the microphone. For voiceovers, a consistent distance and a quiet room usually help more than aggressive processing.

What do the microphone processing options change?

Echo cancellation, noise suppression, and auto gain start enabled. Echo cancellation requests processing intended to reduce sound returning from speakers into the microphone. Headphones can help prevent that feedback during an interview. Noise suppression requests filtering of unwanted background sound. It can make speech clearer near fans or traffic, but may also affect soft syllables, breaths, or the natural texture of a voice.

Auto gain asks the browser to adjust recording level automatically. This can help when speakers vary their distance, though it may also raise quiet background noise between words. For controlled voiceover work, compare short takes with each option enabled and disabled. These settings are browser constraints, so actual behavior depends on device support. Change them before starting; controls stay locked during capture so the take does not unexpectedly switch inputs or processing mid-sentence.

Which download format should you choose?

The original download preserves the compressed audio produced by your browser. The recorder checks support in this order: Opus in WebM, MP4 audio, and Opus in OGG. MP4 recordings use an M4A filename. Browser families and operating systems differ, so the extension can vary between devices. If none of these recording types is available, try a browser with microphone recording support. Renaming an extension does not convert the audio inside the file.

WAV is the option for a trimmed selection or for a workflow that accepts uncompressed PCM audio. After stopping, drag the waveform handles to exclude silence at the start or end. Preview your selected range and leave a little breathing room around speech. WAV exports use 16-bit samples and can be larger than the compressed original. They do not restore detail lost during recording compression. For removing a passage in the middle or adding fades, open your saved take in Audio Cutter.

How do you trim precisely without a mouse?

After the waveform appears, focus a selection handle with the keyboard. Left or Down moves that boundary earlier by 0.1 seconds; Right or Up moves it later by 0.1 seconds. Hold Shift for 1-second steps. Home requests the beginning and End requests the end, with movement constrained so the two handles cannot cross. The editor keeps at least 0.05 seconds between handles, or the whole duration when the take is shorter than that.

Press Space while the waveform editor is focused to toggle preview, or use Play selection and Pause preview. Listen to both ends before exporting so a boundary does not cut off a consonant. Trimming selects one continuous range; it does not join several separate passages. Export rounds each boundary to the nearest decoded sample frame, and Reset trim restores the full decoded duration without making you record again.

What happens when a recording is too long or cannot be decoded?

The recorder does not impose a fixed minutes limit. It accumulates compressed chunks in memory and decodes the finished take for editing. The decoding guard estimates retained floating-point audio as duration × sample rate × channels × 4 bytes. Its budget is 256 MiB when the browser reports device memory of 4 GB or less, and 1,024 MiB otherwise; an unavailable report uses the larger budget. This check runs after browser decoding, so it does not prevent every memory allocation or guarantee a maximum safe session length.

If waveform preparation fails, download the original before trying another take. Preview and WAV export need decoded audio, but the original blob is retained independently. If WAV export fails after a successful decode, try a shorter selection as the error message suggests. For a microphone error, check site permissions, the device connection, and whether another application is using the input. Permission denial and a missing microphone have separate messages. None of these errors means an unsaved take has been backed up.

How does the recorder handle privacy and unsaved takes?

Recording, waveform preparation, playback, and WAV encoding happen on your device. The tool makes no audio uploads and keeps no recording draft in browser storage. Nothing is stored by the recorder after you close the tab. A downloaded file remains wherever your browser saves it, and a share action sends the chosen file to the destination you explicitly select. Those saved or shared copies have their own lifecycle outside this recorder.

A browser leave-page warning is requested while capture is active, including when paused. Browsers decide whether to display that warning, so do not rely on it as a backup. Download the finished take before navigating away. New recording asks for confirmation if the current take has not been downloaded. Stopping capture releases all microphone tracks and closes its analysis context; leaving the page releases recording and playback resources as well.

Share appears only when the browser reports support for sharing files of the relevant audio type. With no trim, it shares the compressed original; with a trim, it prepares a WAV of the selection. Sharing does not mark the take as downloaded, so New recording can still ask whether to discard it. Changing a selection also resets the downloaded flag. Check that your local file or chosen destination has the audio you need before discarding the working take.