Voice Library contains supplied voices you can preview and use. My Voices contains personal voices you have cloned. Both live on the Voices tab of My Assets. Selecting a voice only attaches it to the current AI draft; it does not generate narration by itself.
What do I need?
- Decide whether to use a supplied voice, an existing voice under My Voices, or a new voice clone. Previewing supplied voices needs no upload.
- Voice cloning needs a clean recording containing only the target speaker; record it in ChatCut or upload a file.
- Clone only your own voice or a voice you have explicit permission to use.
Find and preview a voice
- Select + below the AI composer, choose My Assets, and open the Voices tab.
- Choose Featured for supplied voices or My Voices for your clones. Use Search voices and the filters to narrow the list.
- Use the play button on a card to preview the voice.
- Select Use on its card to attach it to the current AI draft.
You can also manage your voices outside a project: open My Assets from the sidebar on the ChatCut Desktop home screen or in the web app, then choose Voices. To give a voice to an external agent that has the ChatCut plugin, select Use in other AI agents on its card and paste the copied prompt into that agent.
Use the language filter to match auditions to the narration you intend to create. Japanese is available for compatible voices: selecting Japanese shows Japanese samples and carries Japanese as the explicit synthesis language. A voice can support some languages but not others, so keep the filter aligned with the script rather than relying on the conversation language.
Clone your voice
- Select + → My Assets, open Voices, switch to My Voices, and select Clone Voice.
- Confirm the voice-rights notice and choose the recording language.
- Select Start recording and read the displayed text, or select Upload audio. Provide 10 seconds to 3 minutes; at least 30 seconds is recommended; if your recording is longer, trim the audio reference before uploading.
- Keep the window open while ChatCut creates a preview, then compare Cloned voice preview with Original audio.
- If needed, edit the preview text and select Regenerate preview. You can regenerate the preview once for free.
- Enter a voice name and select Save. The result appears under My Voices.
Reference audio supports AAC, FLAC, M4A, MP3, OGG, WAV, and WebM up to 25 MB. Background music, multiple speakers, low volume, or the wrong microphone reduce quality, so prepare a clean sample before starting voice cloning.
Use a voice in a request
Narration is generated when you send a draft that carries the voice together with the script, language, delivery, and placement — see AI Voiceover for that request and for prompts you can copy.
In ChatCut Desktop 0.5.5 or later, the Desktop Agent can also create an authorized clone from the complete audio track of a local audio or video asset. It asks for consent and applies the same 10-second to 3-minute and 25 MB reference limits; this Agent workflow does not add trimming controls to the manual cloning window.
Any signed-in account can create a clone of its own voice; the Free plan keeps one saved voice, and generating narration with My Voices requires an active paid membership. The number of saved-voice slots depends on your plan; see Pricing. The cloning window and upgrade screen are authoritative for the account’s live limit.
Fix a short spoken mistake
Ask the ChatCut Agent to fix a specific word or short phrase in an existing recording. The Agent first lets you audition compatible saved voices for that speaker. If none match, reply that you want to continue and it can create an authorized clone from clear single-speaker audio already in the project. The correction replaces only the mistaken phrase and keeps the surrounding take and timing intact.
Only clone a voice you own or have explicit permission to use. For the cleanest result, identify the exact phrase and its replacement, and make sure the project contains a clear sample from the same speaker.
Voices, Skills, Design Styles, and Characters and Products are the reusable context you attach from the composer. For what narration costs, see Generation Credits.