Tools

Open Tools from the sidebar. There are three tools: Clean up, Change voice and Pronunciation.

Clean up a recording

Rescue a noisy voice memo, a Zoom recording or a clip you recorded in a hurry.

1. Drop an audio or video file onto the card, or choose one of your recent takes.
2. Choose what to fix:
   • Reduce noise: removes steady hiss, hum, fan noise and low rumble.
   • Tighten pauses: shortens silences longer than 0.6 s and cuts dead air at the start and end.
   • Even out loudness: brings the clip to a standard level so it sits well next to other clips.
     • Podcast: −16 LUFS
     • Streaming: −14 LUFS
     • Broadcast: −23 LUFS
     • Peaks are always kept under −1.5 dB.
3. Click Clean up.

Compare Before and After by playing them side by side. The meters show how the average level and peak changed, and how much dead air was removed. The cleaned clip is saved to History, and your original is never changed.

Clean up uses the Whisper runtime to read media, so download any Whisper model from Models first.

Change the voice of a recording

Keep exactly what was said, and how it was said (the words, timing, pauses and emphasis), but in a different voice. This is useful for:

• re-voicing a scratch narration in your cloned voice
• making a consistent voice across clips recorded on different days
• anonymising a speaker

1. Drop a recording (up to 15 minutes) or choose a recent take.
2. Under Speak it as, pick Chatterbox Default or one of your voices. To add one, click Clone a voice.
3. Click Change voice. Long recordings are processed in sections split at natural pauses, so no word is cut in half.

This needs the Chatterbox engine. Only convert recordings you have the right to use. Like everything Chatterbox makes, the result carries an inaudible watermark that marks it as AI-generated.

Teach voices how to say words

Voices sometimes mispronounce names, acronyms and brand words. Add a word once, and every voice says it the way you want everywhere:

• Studio
• long renders
• batches and watch folders
• dubs
• stories and audiobooks

1. Under Written as, type the word as it appears in your text, for example Nguyen.
2. Under Say it like, spell it the way it sounds, for example Win.
3. Click Add, or press Return.

Tips:

• Matches are whole words, and capitalization is ignored. Click Aa on a word to match its exact capitalization only. This is useful for IT (the department) versus it.
• Phrases work too: New York → Noo Yawk. Longer entries win over shorter ones.
• Acronyms read letter by letter work best with spaces: AWS → A W S.
• Click a word or its pronunciation to edit it. The Try it box shows exactly what the voices will read.

Your scripts, takes and transcripts keep what you wrote. Only the audio changes.