The free AI voice recording cleaner that also amplifies your voice
Rev Clear is an AI audio cleanup tool built for spoken word: podcasts, sermons, interviews, voice memos, lectures, Zoom and phone recordings, YouTube voice-overs and audiobook narration. A neural network isolates the voice, then Rev Clear repairs distortion, reduces echo and amplifies the voice to clear, consistent loudness — all inside your browser, with nothing uploaded.
What Rev Clear fixes
- AI voice isolationDeepFilterNet3, a published speech-enhancement neural network, runs on your device and separates the voice from noise frame by frame — including changing noise like keyboards, clatter, wind and crowd chatter.
- Transcribe voice to textWhisper speech recognition runs on your device and turns the recording into editable text with clickable timestamps, Voice 1 / Voice 2 speaker labels, an accuracy guard that flags AI guesses, and slow-down playback (pitch kept) for checking words. Copy it, or download TXT, SRT subtitles or VTT captions. 20+ languages, plus translate-to-English. Transcribing the cleaned audio catches quiet and far-away voices a raw recording misses.
- Voice & background separationSplit a recording into a voice-only track and a background-only track (music, room, crowd), or blend some background back under the clean voice.
- Built-in audio editorCut, copy, paste, delete, trim, silence, fades, gain and normalize with zoom and unlimited-feel undo — plus AI tools that isolate or remove the voice in just the part you select.
- Echo & room reductionShrinks the reverb tail of churches, halls and empty rooms so the voice sounds closer to the mic.
- Background noise removalClassic spectral cleanup learns the fingerprint of your room noise — or a noise sample you select — and removes hiss, fans, AC and road drone.
- Distortion & clipping repairFinds waveform peaks that were chopped flat when the mic was too hot and rebuilds them with smooth curves. A “soften harsh distortion” option tames crunchy overdrive.
- Volume boostOne slider (or tap Loud / Very loud / Maximum) makes the cleaned recording much louder — a look-ahead limiter keeps it from ever clipping.
- Voice amplificationVoice Lift raises quiet or far-away speakers, Amplify adds extra volume, and a look-ahead limiter makes it loud without clipping.
- Hum removalAuto-detects 50 Hz or 60 Hz electrical hum and notches out the hum and its harmonics.
- Clicks, pops & crackleRemoves short impulsive clicks from old tapes, vinyl transfers and bad cables.
- Clarity & de-essingCuts muddy boxiness, adds presence so words are easy to understand, and smooths harsh “S” sounds.
- Loudness standardsNormalizes to −16 LUFS for podcasts, −14 for YouTube and Spotify, or −23 for broadcast.
- A/B compare & spectrum viewFlip between original and cleaned instantly with level matching, and see the noise disappear on the spectrogram.
- Instant preview & pause trimmingHear a 15-second preview before cleaning the whole file, and optionally shorten long pauses when you export.
How to clean up a voice recording
- Add audio — drop a file, choose several for batch processing, or record with the built-in recorder.
- Pick a preset — 🔊 MAX VOICE (when voices can’t be heard), Podcast, Sermon / Church, Echoey Room, Heavy Noise, Phone / Voice Memo, Quiet Speaker, Distorted, Old Tape, or Gentle for music. Tap Preview 15 s to hear it first.
- Press Clean Recording and listen — use A/B to compare.
- Edit (optional) — open the result in the built-in editor to cut mistakes, trim, fade or fix one section.
- Download as WAV or MP3, or share straight from your phone.
Who it’s for
Pastors and churches cleaning sermon recordings, podcasters, YouTubers, teachers recording lessons, journalists with field interviews, musicians cleaning vocal demos, and anyone with a voice memo that’s too noisy or too quiet to hear.
Frequently asked questions
Is Rev Clear really free and private?
Yes. All processing runs inside your browser on your own device. Your recording is never uploaded to a server.
Can it fix distorted or clipped audio?
It repairs hard clipping by detecting flattened peaks and rebuilding them, and can soften harsh distortion. Heavy distortion that changed the character of the sound can’t be fully undone by any tool, but it will be noticeably smoother.
My recording is so quiet nobody can hear the voices. What do I do?
Choose the 🔊 MAX VOICE preset. It brings whispers and far-away speakers up to a healthy level before cleaning, lifts every quiet voice up to match the loudest one, and delivers a loud, clip-free recording. If you still need more, tap Very loud or Maximum under Volume after cleaning.
Can it make a quiet voice louder?
Yes. Voice Lift raises quiet or distant speech toward the level of the strongest speech without lifting noise between words. Amplify and the loudness target raise the whole recording, and the limiter prevents clipping.
What background noise can it remove?
With AI voice isolation on, both steady noise (hiss, fans, AC, traffic drone, hum) and changing noise (keyboard clicks, clatter, wind gusts, crowd chatter) are removed, because the network keeps what sounds like the main voice. Background music is removed too — turn AI off for singing or music. Another person talking at the same time may be partly kept.
Can it separate my voice from background music?
Yes. Turn on “Also make a background-only track” to download the music/room without the voice, or use “Remove voice” in the editor on a selection. The voice track is very clean; the background track can keep faint traces of the voice in loud passages.
Can I edit the audio too?
Yes. Open the original or the cleaned audio in the built-in editor to cut out mistakes, trim, fade, adjust volume, and isolate or remove the voice in any selection — then export WAV or MP3.
Can it transcribe my recording to text?
Yes. Tap 📝 Transcribe in step 4. Whisper speech recognition runs on your device — the ≈ 53 MB transcriber downloads once, then works offline, and your audio never leaves the device. You get editable text with timestamps (tap a time to hear that line), and can copy it or download TXT, SRT subtitles or VTT captions. Choose the cleaned recording for the best accuracy. Speed depends on the device: a modern laptop is usually faster than real time, phones are slower. Like every automatic transcriber it can mishear names and unusual words, so read it over before publishing.
Can it clean and transcribe security camera or doorbell audio?
Yes. Pick 🎥 Far voices / camera: it keeps faraway voices that full-strength AI noise removal would treat as noise, and lifts them. In the transcript, two people are labelled Voice 1 (lower) and Voice 2 (higher) by their voice pitch — tap a label to fix it. Lines the AI is only guessing (like a lone “Thank you.”) are marked unclear? instead of being passed off as real speech, and you can replay any line at 0.75× or 0.5× with the pitch kept to type what you hear. Speech that is very far from the mic can still be impossible to recover word for word — the app tells you honestly when that happens.
Is the AI private? Does it need the internet?
The AI runs inside your browser. Its model (about 12.6 MB) downloads once from this site and is stored for offline use. Your audio never leaves your device.
How fast is AI cleaning?
On a typical laptop the AI runs several times faster than real time; phones and older computers are slower. Use Preview to try settings in seconds before cleaning a long file.
Can it remove echo?
Partly. The AI already reduces room sound, and the Echo / room reduction control removes more of the reverb tail. Very strong echo can’t be removed completely by any tool.
Why does my cleaned audio sound “watery” or robotic?
That’s over-processing. Lower Noise Reduction or Noise Detection Strength, or try the Gentle preset. Leave some natural room sound for the most natural voice.
What file types are supported?
Anything your browser can decode: MP3, WAV, M4A/AAC, OGG, OPUS, WEBM, and FLAC on most browsers, plus the audio track of many MP4/MOV videos. Export as WAV (16 or 24-bit) or MP3.
How long can my recording be?
Long recordings work, limited by your device’s memory. An hour of audio is fine on most laptops; on older phones, recordings under about 30 minutes are most reliable.
Does it work on my phone?
Yes. It runs in mobile browsers and can be installed to your home screen. Once installed it also works offline.