/voice — AI Voice Transformation
Transform any voice with AI The /voice command re voices an audio recording: the words, timing, and delivery stay the same, but the speaker's voice is replaced with the one you pick. It runs on Ele...
Transform any voice with AI
The /voice command re-voices an audio recording: the words, timing, and delivery stay the same, but the speaker’s voice is replaced with the one you pick. It runs on ElevenLabs Voice Changer and returns an MP3.
How to use
- Attach an audio file with speech (MP3, WAV, M4A, and other common formats)
- Type
/voice - Pick a target voice — a preset or a custom ElevenLabs Voice ID
- Optionally turn on Remove background noise
- Send to transform
/voice [with an audio file attached]
Voice options
Default voices — the same 20 voices as /tts, each with a play button to preview before you commit:
- Female: Aria (default), Sarah, Laura, Charlotte, Alice, Matilda, Jessica, Lily
- Male: Roger, Charlie, George, Callum, River, Liam, Will, Eric, Chris, Brian, Daniel, Bill
Custom ID — paste any voice ID from the ElevenLabs Voice Library to transform into a voice outside the preset list.
| Option | Description |
|---|---|
| Target voice | One of the 20 presets, or a custom ElevenLabs Voice ID |
| Remove background noise | Cleans noise out of the recording while transforming |
Use cases
- Content creation — voiceovers in different voices without re-recording
- Privacy — anonymize the speaker in a recording
- Podcasting and fiction — consistent character voices
- Narration — match a specific vocal style across clips
Credits
/voice bills per second of audio at 60 credits per minute, measured on the finished file:
| Clip length | Cost |
|---|---|
| Up to ~5 seconds | 5 credits (minimum charge) |
| 1 minute | 60 credits |
| 5 minutes | 300 credits |
| 10 minutes (maximum) | 600 credits |
The picker shows 5 credits before you attach a file — that is the minimum charge. Krater reserves an estimate up front and settles the real amount from the generated audio, so over-estimates are refunded automatically. Failed transformations are refunded in full.
Limits and requirements
- Up to 10 minutes of audio per transformation
- Requires a paid plan (Pro, Ultra, or Max) or a team seat; not available on free accounts
Tips
- Clean, close-mic recordings transform best; turn on Remove background noise if yours is not
- Preview a few voices first — different voices handle accents and pitch differently
- For text-to-speech instead of re-voicing a recording, use /tts; to turn speech into text, use /transcribe; to talk to Krater live, see Voice Mode
More in Slash Commands
- /image — Generate & Edit Images
- /video — Create Videos
- /upscale — Upscale Images & Video
- /build — Generate Apps, Sites & Code
- /sound — Music & Sound Effects
- /tts — Text to Speech
- /transcribe — Transcribe Audio & Video
- /research — Research
- /summarize - Summarize URLs, Videos & Files
- /detect — AI Detection
- /plagiarism — Check Plagiarism
- /humanize - Make AI Text Natural
- /compare — Compare AI Models
- /slides — Generate Presentations
- /document — Create Documents
- /3d — Generate 3D Models