Cleanvoice

New

Browser tool that removes filler words, stutters, mouth sounds, long silences and background noise from podcast audio, with multitrack sync, timeline export to DAWs, transcripts and an API.

cleanvoice.ai

Cleanvoice

Explore features, practical uses and pricing below.

Cleanvoice is a browser tool for spoken audio with one main job: it finds the ums, stutters, lip smacks and dead air in a recording and takes them out. It can also reduce background noise and enhance the voice. You don't edit on a timeline inside Cleanvoice. You upload, choose what to remove, and either download the cleaned file or export the edit to your own audio editor to review each cut. It runs in desktop and phone browsers, with no software to install.

What Cleanvoice removes from a recording

Each cleanup tool is listed separately on the site:

  • Filler Words Remover for "um", "uh" and similar sounds. The FAQ says it supports more than 20 languages.
  • Stutter Remover for repeated word starts.
  • Mouth Sound Remover, with breath removal, for the clicks and lip noise close microphones pick up.
  • Deadair Remover to shorten long silences.
  • Background Noise Remover and Studio Sound, an audio enhancer the company says works in any language because it operates at the phonetic level.

Timeline export, multitrack and video podcasts

For careful editors, Timeline Export is the feature that matters most. Instead of accepting a rendered file, you can open the edit in Adobe Audition, REAPER, Audacity, DaVinci Resolve or Ultraschall and check or reverse individual cuts. Multitrack editing processes several speaker tracks and keeps them in sync. Video podcasts are supported, though the site is clear that Cleanvoice changes only the sound, and visuals stay in your video editor. Transcription and a summary with show notes come with every plan.

The API, SDKs and automation

The Cleanvoice API documentation describes a v2 REST API: you submit an edit, then poll its status until processing finishes. Python and JavaScript SDKs handle that polling, and Make.com and n8n cover no-code automation. A separate Speech Enhancement SDK and API is offered, as is a Cleanvoice MCP for connecting the service to AI assistants. That makes it usable as one step in a pipeline, for example cleaning every new recording that lands in shared storage.

Podcasters it suits, and who needs more

Interview podcasters who talk naturally and leave dozens of fillers per episode are the clearest fit, along with freelance editors who want a first pass done before they start. Shows recorded in languages other than English benefit from the multilingual filler detection. Cleanvoice doesn't record and has no script or transcript editor, so if you want to restructure an episode by editing text, Descript or Async covers more of the job.

Cleaning a two-person interview

  1. Upload the host and guest tracks as a multitrack project.
  2. Turn on filler, stutter and mouth-sound removal, and dead-air removal for long pauses.
  3. Leave Studio Sound off at first if the guest's mic already sounds good, and compare the two versions.
  4. Export the timeline to REAPER or Audition, scrub through the cuts, and restore any that removed a meaningful hesitation.
  5. Mix the episode, then use the transcript and show notes as a draft for the episode page.

What automatic removal can get wrong

Fillers sometimes carry meaning, as in "I, um, wasn't sure", and removing every one can make a speaker sound rushed or rehearsed. Cutting breaths too tightly can create joins that sound unnatural. That's why the timeline export matters: plan time to review rather than publishing the rendered file unheard. Strong accents and crosstalk can affect detection, so use the free trial on your own show's audio instead of a clean sample.

Trial, credits and subscriptions

New users can clean 30 minutes free with no credit card, according to the pricing page. After that you buy pay-as-you-go credits, which the page says stay valid for two years, or a monthly subscription where unused credits roll over up to three times the plan limit. A custom plan covers higher volumes and custom API endpoints, and the page states that all plans include all features.

Cleanvoice questions

Can I undo a cut I disagree with?

Yes, by exporting the timeline to a supported editor and restoring the cut there.

Does it change my video?

No. Cleanvoice edits the audio of a video podcast, and the picture edit stays in your video software.

Similar tools

Record podcasts and video interviews with local multitrack capture up to 4K, then edit by transcript, clean audio with Magic Audio, cut short clips, translate, livestream and publish from one platform.
Web service that levels speech and music, normalizes loudness to set targets, reduces noise and reverb, cuts filler words and silence, and adds transcripts, chapters and publishing for podcasts.
Turn supplied music or audio into waveform videos with custom artwork, text, visual effects and automated lyric visualization.