Repair what the camera and mic got wrong.

Most bad recordings share the same problems: a laptop fan, an air conditioner, a street outside, a bare room that rings. Cleanoise separates the speech from everything else and lowers the rest, so the voice stays in front.

Echo is handled separately from steady noise. A hard-walled room and a humming fridge need different treatment, so you can set each strength on its own instead of accepting one global setting.

There are limits. If the speech itself is clipped or the room noise is louder than the voice, no tool can fully recover it. We show a before/after preview first so you can judge the result before you spend a credit on the export.

When you read a script or watch your own face, your gaze drifts below the lens. Eye-contact correction adjusts how your eyes appear so you look toward the camera again.

It is optional and off by default, including in the Fix all preset. You choose when to use it, and you control the strength. It adjusts the appearance of your own gaze; it does not swap in another face or change who is speaking.

It works best on a mostly front-facing face with good light. Extreme angles, heavy blur and glasses glare reduce the quality, and we will tell you when a clip is a poor candidate.

A window behind you, a lamp to one side, a warm ceiling bulb: webcams struggle with all of it. Cleanoise measures the frame, lifts a dim face and evens out the exposure without blowing out the highlights.

Color balance fixes the orange or green cast that comes from mixed lighting, so skin tones look natural and white things look white.

Both adjustments have a strength slider. Light correction is a repair, not a replacement for a lamp: a very dark clip will gain noise when brightened, and we say so in the preview.

Phone and laptop microphones often sound thin or boxy. Speech clarity brings out the consonants and reduces the muddy low-mid buildup that makes a voice hard to follow.

Loudness normalization brings the whole video to a consistent level, so viewers are not reaching for the volume control between a quiet intro and a loud answer.

This improves what was recorded; it cannot restore frequencies a microphone never captured. Clipped (distorted) peaks are the main thing we cannot undo.

If you do not want to think about settings, Fix all applies a balanced combination of the fixes that are currently available: noise and echo, lighting, color and speech.

Eye-contact correction is never included automatically. If and when it is available, you turn it on yourself.

The preset is a starting point. After it runs you can open each fix, change its strength and re-preview before exporting.

Course creators and teams often have dozens of recordings with the same problems. Batch processing applies one preset to every file in a queue.

You would review a sample first, then run the rest. Results land in a single export folder so you are not downloading file by file.

Batch is on the roadmap and not yet available. It is listed on the pricing page roadmap and is not part of any plan today.