Getting started with speech denoise
Your recordings and results are linked to your account. Files are uploaded for processing; AI tools also send audio to an external processing service.
01Choose the right tool
Use noise reduction for hiss, fans or steady room noise in speech. Smart cleanup is useful for more varied background sounds; basic cleanup keeps processing on our server. Video cleanup processes the soundtrack and exports a supported video as MP4. Vocal separation makes two tracks from music: vocals and accompaniment. Speech enhancement is for difficult spoken recordings; it cannot guarantee removal of strong echo.
02Listen before uploading
Each tab has a one-click example with an original and a saved result. Listening does not require sign-in or spend credits. These are examples, not a promise that every recording will sound the same. For separation, compare both tracks: some instruments may leak into vocals and some singing may remain in the backing track.
03Upload and check the recording
Sign in with Google when you are ready to process your own file. Choose the matching tab, select a supported file and listen to its original sound. The form shows the current size and duration limits. Use the highest-quality original available; converting a compressed file to WAV does not restore missing detail. Video must include an audio track.
04Review cost, then start once
The estimate uses the selected option and rounds the duration up to a started minute. The server checks the actual duration and available credits before processing. An insufficient balance opens the pricing dialog. Wait for the current task to finish; if its status is uncertain, recover that task from history instead of submitting a new one.
05Compare the result fairly
Listen to the same section before and after at a similar volume, preferably with headphones. Check quiet words, consonants and pauses as well as loud noise. Reducing noise too strongly can make speech thin or metallic. A distorted, clipped or heavily reverberant source may remain imperfect. Keep the original for comparison.
06Edit, export and find it later
After processing, set an export name, choose a start and end time and optionally adjust volume gain. Save changes before downloading the saved version. MP3 is compact; WAV suits further editing; supported video results can be exported as MP4. Editing and exporting an existing completed task do not run cleanup again. Your personal center lists tasks from every tool.
07Understand uploads and deletion
All processing uploads your recording to our server. Smart processing also sends audio to an external processing service. Basic noise reduction stays on our server. Results are private to your signed-in account. Delete recordings you no longer need and retain your own backup; the service is not permanent file storage.
08If something goes wrong
If a file is rejected, check its format, size, duration and audio track. For a pending task, use recovery for the existing task. A confirmed processing failure returns the task credits once; an uncertain external status keeps them reserved. If playback fails, retry after checking your network. Contact account support with the task ID, without sending passwords or payment details.