# Custom Voices

Custom Voices let you create your own voice options to use with Text-to-Speech. Instead of being limited to the default voice library, you can upload audio samples, generate a unique voice from them, and attach it to any of your characters.

Once a Custom Voice is set as a character's voice, anyone playing TTS on that character will hear your voice — it becomes part of the character's identity.

{% hint style="info" %}
Custom Voices are an **I'm All In** subscriber feature. 
{% endhint %}

### Creating a Custom Voice

Creating a Custom Voice is a three-step process: open the creation form, fill in the details and upload your audio, then let it process.

#### Step 1: Open the Custom Voice form

1. Open the sidebar menu.
2. Expand the **Create** section.
3. Click **Voice**.

This opens the Custom Voice creation form.

#### Step 2: Fill in the details

On the form, complete the following fields:

1. **Name** — Give your voice a memorable name so you can find it later (up to 20 characters).
2. **Description** — Add a short description of the voice, such as its tone or mood (up to 100 characters).
3. **Language** — Select the language your audio samples are recorded in.
4. **Attach voice** — Click the upload area or drag and drop your audio file(s).

{% hint style="info" %}
**Upload requirements**

* MP3 or WAV format only
* Between 5 and 30 seconds of continuous speech per clip
* Up to 3 clips per voice
* All files combined must be under 4 MB 
  {% endhint %}

#### Step 3: Create the voice

Once all the fields are complete, click **Create Voice**.

Your voice will then be sent for processing. This can take a while, so don't worry if it isn't ready immediately.

### Previewing and Saving Your Voice

Once processing is complete, you can preview the result before deciding to keep it.

1. Go to **My Creations**.
2. Open the **Voices** tab.
3. Find your new voice and press the preview button to listen to it.
4. If you're happy with the result, save it.

Your saved voice is now ready to be attached to a character.

{% hint style="info" %}
You can have up to **20 Custom Voices** on your account at any time.
{% endhint %}

### Adding a Custom Voice to a Character

A Custom Voice doesn't do anything on its own — it has to be attached to a character. Once attached, it becomes part of that character's identity and will play whenever someone uses Text-to-Speech with the character.

1. Create a new character, or edit an existing one.
2. On the edit page, scroll down to the **Attach voice** section.
3. Select the Custom Voice you want to use.
4. Save your character.

{% hint style="warning" %}
If your character is **public**, anyone using it will hear your Custom Voice when they use Text-to-Speech. 
{% endhint %}

### Tips for Creating Great Custom Voices

The quality of your source audio directly affects how good your Custom Voice sounds. These best practices will help you get the best possible result.

#### General best practices

* **Choose a script that matches your target voice.** Your recording should reflect the voice you want to generate. If you want an energetic, expressive clone, don't read a flat script in a monotone — pick a script with the right energy and emotion.
* **Record clearly in a quiet space.** Use the best microphone you have access to, and record somewhere quiet. Background noise, echo, or muffled audio will all show up in the final voice.
* **Avoid long pauses.** The cloned voice mimics the pacing of your recording, including pauses between sentences. Keep your delivery natural and continuous.
* **Trim your recording.** The clip should be roughly all speech from start to finish — no silence at the beginning or end, and no cut-off words. Free tools like Audacity make it easy to clean up a recording.
* **Speak in the target language.** If you want your voice to speak English, record in English. Recording in a different language than your intended output will reduce quality.
* **Keep your recording around 10 seconds.** This is the sweet spot for high-similarity clones. Longer clips do not produce better results.
* **Lower the Response Max Token setting in chat.** When using your Custom Voice in chat, setting **Response Max Token** to its minimum value usually gives the best-sounding output.


---

# Agent Instructions: Querying This Documentation

If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question.

Perform an HTTP GET request on the current page URL with the `ask` query parameter:

```
GET https://docs.spicychat.ai/product-guides/premium-features/text-to-speech/custom-voices.md?ask=<question>
```

The question should be specific, self-contained, and written in natural language.
The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.