$ one file in, one file out

CleanTake

Clean recorded speech locally — no upload.

Drop in a noisy voice recording. Get back clean speech. Everything runs on your machine — your audio never touches a server, and there's no monthly tab.

Hear the before / after Download v1.0.1

What it does

Drop a file

WAV or MP3 of already-recorded speech — a podcast take, a voiceover, a lecture.

Local denoise

Fan hum, air-conditioner drone and steady room noise are stripped out on your CPU. Nothing is uploaded.

Get clean audio

Loudness-normalized output lands next to your input. One file in, one file out.

Honest scope — read before you try

Removes steady environmental noise: fans, air-conditioners, electrical hum, room tone.
Works on already-recorded files (podcasts, voiceovers, online lessons).
Does NOT separate overlapping speakers — it can't pull one voice out of a crowded café.
Is NOT a real-time call tool — it processes files, not live meetings.

We publish our limits on the limits page and keep a public changelog. If your use case is outside this scope, CleanTake isn't the tool — and we'd rather tell you now.

Why local-first?

Cloud audio tools charge a monthly subscription and require you to upload your voice to someone else's server. CleanTake does the job on your own machine: private by architecture, free by design. No account, no telemetry (see our privacy note).

Competitor pricing and features change often — anything we compare is re-checked on the day of publishing, never treated as permanent data.

Status

CleanTake v1.0.1 is out (2026-10-02). The core pipeline (ffmpeg → DeepFilterNet → loudness normalization) passed a real-voice test: noise floor −33dB → −59dB on environmental noise, voice preserved. Download the Windows installer or the portable ZIP — free, no account. See the changelog.

Download v1.0.1 — free

From the blog

Deep-dives with real evidence — benchmarks, honest limits, workflows. All posts →