> ## Documentation Index
> Fetch the complete documentation index at: https://docs.wittify.ai/llms.txt
> Use this file to discover all available pages before exploring further.
# Voice Picker
> Browse the full voice catalog and pick a voice for your generation.
**Choose a voice** is the modal that opens when you click the voice card in the Studio's right panel, when you click any voice slot in the agent builder, or from the **+ Create Voice** topbar shortcut. It carries every Wittify voice across four tabs, plus a search field, a Filters popover, a Language and Dialect chip row, and a brand-gradient **Create A Voice** button that routes into the Clone or Design wizard.
## The modal at a glance
The picker is a wide dialog (up to 980 px on a desktop). It always shows the close icon in the top-end-edge corner.
| Region | What you see |
| ------------------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------- |
| Tab bar | Four pill tabs grouped together on the start edge: **Explore**, **My Voices**, **Favorites**, **All Voices**. The active tab gets a card-tinted background. |
| **Create A Voice** button | A brand-gradient pill with a star icon on the end edge of the tab bar. Opens the **Create a new voice** picker. |
| Search bar | Spans the full width below the tabs. The placeholder reads *Search voices...* It searches across name and dialect. The input is capped at 120 characters. |
| **Filters** button | Right after the search bar, with a sliders icon and a small badge showing the count of active filters. |
| **Clear filters** link | Appears only when at least one filter is active or the search box has text. |
| Language chip | A pill below the search row showing the active macro language with a flag and chevron. Default reads **Language**. |
| Dialect chip | Appears next to the language chip when the chosen language has a dialect list. Default reads **Dialect**. |
| Voice list | The body of the modal. Scrolls vertically. |
## The four tabs
The default tab. Shows a header *Get started with these voices*, a one-line subtitle *A selection of voices to try first*, and a curated list of up to 6 voices that pass the active language, dialect, and filter selection.
When no voices match, the empty state reads *No voices match your filters.*
Explore renders **only** the Get Started list. The old "Tailored for your use case" gradient cards block and the use-case chip strip next to the language selector were both removed per product direction. They must not be reintroduced.
Your custom voices. Cloned and designed voices made from your account both surface here. The list also passes through the active filters.
When you have not made any voices yet, the empty state reads *You haven't created any voices yet.* with a brand-gradient **Create your first voice** button below it. Clicking it routes into the same Clone or Design picker as the **Create A Voice** button.
Voices you have starred from any tab. Click the star on a voice row to add or remove it. Favorites are scoped to your account.
When the list is empty, the modal shows the same *No voices match your filters.* fallback.
Every Wittify voice plus your custom voices, all in one list, after passing through the active filters and search.
When filters or search return nothing, the empty state reads *No voices match your filters.*
## The voice row
Every row in the list looks the same regardless of tab. From the start edge to the end edge:
| Element | What it shows or does |
| -------------------------------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| Avatar circle | A 40 px round tile with the voice's first letter and a deterministic color tint generated from the voice id. The same voice always gets the same tint. |
| Voice name | The voice name (Arabic name when available in Arabic mode). |
| Gender pill | A small chip next to the name. **Male** in blue, **Female** in pink. |
| Country flag | A flag emoji (when the voice has a known dialect, hidden on small screens). |
| **Select** / **Selected** button | The action button. Reads **Select** when the voice is not active, **Selected** with a check mark and a brand-gradient background when it is. |
| Play / Stop button | A small square with a play triangle. Click to preview the voice with a built-in 140-character sample line (Arabic or English depending on UI language). The icon flips to a stop square while playing, and to a spinner while loading. |
| Star button | A star icon. Empty when the voice is not in your favorites, gold and filled when it is. |
| Sample line by language | Text |
| ----------------------- | -------------------------------------------------------------------------- |
| English | *Hi there, this is a sample of my voice. I can read anything you send me.* |
| Arabic | *مرحباً، هذا مثال على صوتي. يمكنني قراءة أي نص تقدمه لي.* |
Click the play button on more than one row in a row. The earlier preview cancels automatically, you never end up with two voices speaking at once.
## The Filters popover
Click **Filters**. A small popover opens on the end edge of the modal.
| Section | What it shows | How it works |
| ---------------------- | ------------------------------------------------------------------------ | -------------------------------------- |
| **GENDER** | Three radio chips: **Any**, **Male**, **Female**. | Pick one. **Any** is the default. |
| **Clear filters** link | Appears at the bottom-end edge of the popover when any filter is active. | Resets gender to default in one click. |
A small numeric badge appears on the **Filters** button when gender is active (1 if not **Any**).
The picker does not have a separate Use Case filter. Use cases live inside the Clone Wizard's Tags step, they are not a discovery filter on this surface.
## The Language and Dialect chips
These are quick-narrow pills below the search row.
### Language
The default value reads **Language**, with the 🌐 globe glyph. Click to open a popover with:
| Inside the popover | Notes |
| ------------------ | -------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| Search input | Filters the list by English name, Arabic name, or language code. Up to 200 rows render at once. |
| **Any** row | First in the list, with the 🌐 glyph. Resets the language filter. |
| Per-language rows | Each shows a country flag, the language name (English or Arabic depending on UI language), and a small **+** badge on the end edge for languages with dialect lists. |
Selecting a language resets the dialect to **Any**.
### Dialect
The dialect chip only appears when the chosen language has a dialect list. Click to open a popover with:
| Inside the popover | Notes |
| ------------------ | -------------------------------------------------- |
| **Any** row | Default. Uses the parent language's flag. |
| Per-dialect rows | Each shows the regional flag and the dialect name. |
The catalog covers 51 clusters: Arabic (21 varieties), English (10 accents), Chinese (15 varieties), and 48 more clusters with 2 to 17 dialects each. Arabic and English are first in the macro list per product direction.
The Language and Dialect chips on the picker only narrow which voices show up. They do not change the language the voice will actually speak, that is set on the Studio's right column or in the agent builder.
## Picking a voice
Click **Select** on any row. The picker closes and the chosen voice replaces the previous one in the calling surface (Studio's voice card, the agent builder's voice slot, etc.). If you click **Select** on the voice that is already active, nothing changes, the picker just closes.
## Creating a voice
Click **Create A Voice** on the tab bar (or **Create your first voice** in the empty My Voices tab). The picker closes and the **Create a new voice** dialog opens.
| Card | Title | Description |
| ------ | ------------------ | --------------------------------------------------------------------------------------------------------------------------------------------- |
| Clone | **Clone a voice** | *Upload or record a short sample. We will match tone, accent, and pacing.* Opens the [Clone Wizard](/en/Systems/Text-to-Speech/Clone-Wizard). |
| Design | **Design a voice** | *Describe how it should sound, and we will generate it from scratch.* Opens the [Design Wizard](/en/Systems/Text-to-Speech/Design-Wizard). |
When the wizard finishes, the new voice is auto-selected on the calling surface and shows up in **My Voices** the next time you open this picker.
## Provider routing
Voices are Wittify voices. The underlying provider is internal routing detail and never appears anywhere on a row. The avatar tint, the gender chip, and the country flag come from the voice itself, not from the provider.
The picker must never display provider names. The avatar's color tint is derived from the voice id, the gender chip is derived from the voice's gender field, and the flag (when present) is derived from the dialect, not the provider routing.
## Common questions
The preview request is taking longer than usual. Click another row and the spinner cancels. If preview consistently hangs, the underlying voice provider is having a moment and a generation in Studio will likely fail too.
Look in the **My Voices** tab. If the voice is still in **Training**, it shows up but cannot be picked yet. Voices in **Failed** state still appear. Voices you have not made are not in **My Voices**, they are in **All Voices**.
They are independent. **Filters** narrows by gender. **Language** and **Dialect** narrow by what the voice supports. A voice has to pass all three to appear in a tab.
**I see voices listed in **All Voices** that did not show up under **Explore**.**
**Explore** is a curated short list (up to six voices). **All Voices** is the full catalog. If you want to see everything, go to **All Voices**.
The search matches name, Arabic name, and dialect code, all together. Add more specific words to narrow.
Not from this picker. The voice picker is a tool for choosing a voice on your own session. Cross-account voice sharing is not supported today.
A voice with **Multilingual** support always passes the language filter. That is intentional — multilingual voices read most languages.
## Where to go next
Use the picked voice to generate speech.
See and manage your custom voices.
Add a cloned voice from a sample.
Add a designed voice from attributes.