Microphone selection
Choose the built-in microphone, headset or external input that should be recorded.
Select a microphone, watch the input level and capture a voice memo, narration or test take. Pause and continue the same recording, listen back and download a separate audio file.
Show local-processing and no-upload copy only after confirming how the recorder, encoders, diagnostics and analytics handle microphone data.
Read the full guideInspect the complete path: getUserMedia, MediaRecorder or custom buffers, any WASM encoder, error logging, analytics and download creation. Only say the recording stays on the device when no audio payload is uploaded. Stop all tracks after recording and disclose any persistent storage.
Voice Recorder captures microphone audio directly in your browser so you can make quick voice notes, narration takes, pronunciation samples, support messages, or QA clips without uploading private audio to a server.
Use Test Mic to grant permission and confirm the selected device. Start Recording begins a local MediaRecorder session. Pause and Resume keep one continuous clip, while Stop finalizes it for preview.
MP3 is the practical default for sharing, WAV is best when you need a lossless edit master, and M4A is useful for Apple workflows. The app converts after recording, so the original browser capture stays available until you delete it or close the tab.
The recording, preview, and export process runs in this browser session. Files are not uploaded, and device tracks are released when you leave the page or start a new setup.
Need the cleanest result? Use a headset, record in a quiet room, keep mono enabled for speech, and use WAV if another editor will process the audio later.
A browser voice recorder captures the microphone selected for this tab, keeps one take running across pause and resume, then creates a downloadable audio file. The available containers and codecs must come from the implementation rather than a fixed marketing list.
Choose the built-in microphone, headset or external input that should be recorded.
Temporarily stop adding audio without creating a separate take, when the recorder state supports it.
Request browser echo cancellation, noise suppression or automatic gain only when those MediaTrack constraints are actually wired.
Listen to the completed take, check duration and file size, rename it and save the result before leaving the page.
Select the intended input and speak briefly to confirm the meter responds.
Pick only formats and quality controls exposed by the current engine. Decide whether browser input processing should be requested.
Press Record and keep the page active. Pause for interruptions and resume when ready.
Finalize the take, listen from beginning to end and check for silence, clipping, interruptions and the correct microphone.
Choose a file name and save the recording. Object URLs and in-memory buffers may disappear when the page is refreshed.
A file extension does not guarantee the same codec in every browser. Populate the list from formats that the recorder and export path can actually produce.
Higher lossy bitrate can retain more detail but produces a larger file. It cannot repair a noisy or clipped input.
Mono is usually sufficient for a single voice and saves space. Choose stereo only when the input and use case contain meaningful channel separation.
These are browser capture constraints, not full studio restoration. They can reduce steady noise or speaker echo but may also alter speech.
AGC changes the live input level to keep speech usable. It is different from analyzing and normalizing the completed recording.
Inspect the complete path: getUserMedia, MediaRecorder or custom buffers, any WASM encoder, error logging, analytics and download creation. Only say the recording stays on the device when no audio payload is uploaded. Stop all tracks after recording and disclose any persistent storage.
Verify the meter moves before recording, select the correct microphone and check hardware and system mute controls.
Keep the tab active, avoid disconnecting the device and prevent the phone from locking. Permission changes or memory pressure can end the track.
The browser may have produced a codec that the current preview player cannot decode. Offer only verified export formats and test the downloaded file in a compatible player.
Automatic gain or noise suppression may be reacting to the room. Disable the relevant constraint if the UI and browser allow it, then record again.
Do not require an account unless the product actually gates the tool. The page should state the real access model rather than a generic “free” claim.
Use the formats returned by the recorder and export implementation for the current browser. Do not promise MP3, WAV or M4A universally without verified encoders.
Yes when pause and resume are implemented on the active recording session. The timer and state should clearly show that no second file is being created.
Only claim local processing after verifying that raw audio, encoded chunks and diagnostics never leave the browser. URL-free microphone capture alone does not prove that.
Browser memory, battery, tab suspension and device limits make very long sessions risky. Record a short test and save important takes promptly.
Check microphone distance, system input gain, the selected device and whether automatic gain is disabled. Avoid fixing a very low source only after recording.
Browser capture suppression may reduce some steady noise when supported, but it is not equivalent to dedicated voice cleaning and can introduce artifacts.
Before download, Delete should discard the in-page recording buffer. It does not delete a copy that you already saved.