How to clean up audio
- Choose an audio or video file up to 100 MB and 10 minutes.
- For video, leave Keep the video checked to retain the picture. For audio-only output, choose MP3, WAV or M4A.
- Choose Reduce background noise, Even out volume, or both. Compare the result with the original player before downloading.
Made for dialogue
Try it on speech recorded near a fan or other background noise. Unpin uses RNNoise to reduce noise; it is not a music stem separator, mastering tool or guaranteed echo remover. Voices can sound different after processing. Keep your original.
Even out volume
Even out volume makes quiet and loud speech more consistent, then aims for a steady speech level of about −16 to −18 LUFS and keeps peaks in the saved file below −1.5 dB, so the result does not clip. It works on the cleaned speech and does not change timing, so video stays in sync. Leveling can make any remaining background noise easier to hear; turn on noise reduction too if that happens. A silent or nearly silent recording is left unchanged. This is noise reduction and volume leveling, not studio restoration.
Audio and video support
Inputs: MP3, WAV, M4A, AAC, OGG, MP4, MOV, WebM and MKV, subject to codec support. Cleaned audio is mono. Audio-only exports are MP3, WAV or M4A. The duration limit is 10 minutes, even if the file is smaller than 100 MB.
What happens to a video?
With Keep the video enabled, Unpin copies the first video stream and replaces the audio with cleaned speech. It keeps the source container: MP4, MOV, WebM or MKV. Additional audio tracks and subtitles are not retained. Check synchronization before using the result.
Local processing, honest limits
The processing engine and noise-reduction software download on first use; your recording is processed in this tab. Keep the tab open. A busy background, echo or a clipped microphone may remain audible. If the voice sounds worse, use the original rather than repeatedly filtering it.
Before you start
Will this make my recording sound like a studio podcast?
That is not guaranteed. This tool reduces background noise in speech. It cannot recover every detail missing from a poor recording.
Can I clean a song or isolate vocals?
This speech-focused tool is not designed to separate singing from music. It may alter sounds you wanted to keep.
Why did a small file fail?
A recording can be under 100 MB but longer than 10 minutes. Unsupported codecs or insufficient device memory can also stop processing.