> ## Documentation Index
> Fetch the complete documentation index at: https://easy-peasy.ai/docs/llms.txt
> Use this file to discover all available pages before exploring further.
# Voices and voice cloning
> Find a voice by language and accent, clone your own voice from a recording, or design a new voice from a description.
Every text-to-speech voice comes from the voice list: about 2,000 ready-made voices, plus the voices you add yourself. Your own voices work across Easy-Peasy.AI, for example in [talking videos](/docs/videos/talking-videos) and the [Video Agent](/docs/videos/video-agent).
## Choose a voice
On [AI Text to Speech](/docs/audio/text-to-speech), click **Change voice**.
* **Search voices...** finds voices by name, language, accent, or description.
* **Language** and **Accent** narrow the list, for example English with a Singaporean accent.
* **Audio tags** shows only voices that can act out [audio tags](/docs/audio/text-to-speech#audio-tags).
* The play button plays a sample. The star marks a favourite; favourites come first in this browser.
Click a voice to choose it, then click **Select Voice**. Your own voices appear first in the list.
Two badges to know:
* **Audio tags**: the voice acts out tags like `[laughs]`. Up to 5,000 characters per generation.
* **Saver**: uses one character for every 10 you convert. It always makes MP3 and ignores Turbo and Advanced Settings.
## Your own voices
Add voices three ways:
| Option | How it works | Plans |
| ---------------------- | --------------------------------------------------------------- | ------------------------------------------ |
| **Instant** clone | Copies a voice from a short recording in seconds | All |
| **Professional** clone | Captures more of the speaker's character from a clean recording | Unlimited 50 and higher, Teams, Enterprise |
| **Design a Voice** | Creates a new voice from a written description | All |
All three count toward your plan's custom voices, from 1 on the Free plan. See [Audio plans and limits](/docs/audio/limits). Creating a voice doesn't use characters.
Only clone your own voice, or a voice you have permission to use.
## Instant clone
On AI Text to Speech, click **Clone Your Voice**. **Instant** is selected.
Upload an audio or video file of up to 10 MB, or click the microphone and record up to 3 minutes. Use one speaker, more than a minute of speech, and no background noise or music.
Enter a **Voice Name** and click **Add Voice**. The voice is ready to use right away.
## Professional clone
A Professional clone sounds closer to the speaker, especially their rhythm and expression.
1. Click **Clone Your Voice** and choose **Professional**.
2. Upload an MP3, M4A, or WAV file, or record. The recording must be 10 seconds to 5 minutes long. If it's longer, drag the highlighted window to choose which part to use.
3. Optionally type **What does the recording say?**: the recording is checked against it, which catches silent or unclear samples before the clone is made.
4. Name the voice and save it.
Professional clones:
* Can't be deleted by you. Contact support to remove one.
* Can't be used in the Voice Changer. Make an Instant clone of the same recording for that.
* Ignore **Stability** and **Clarity**, and don't offer Turbo.
* Are removed if your subscription ends.
## Design a voice
Describe the voice you want, and you get three versions to choose from.
Click **Design a Voice** and write a **Description** of 10 to 1,000 characters: age, gender, accent, tone, pace, and the job it's for. **Try an example** fills in a sample, and **Best practices** has tips.
After up to 30 seconds, you get three versions. Play each one.
Select your favourite, enter a **Voice Name**, and click **Save voice**.
Designed voices support audio tags. They appear at the top of the voice list:
## Delete a voice
In the voice list, click the trash icon on one of your voices. Deleting frees a place for a new custom voice. Professional clones can't be deleted this way.
Make audio with your voice.
Custom voices and characters per plan.