# TTS Library - Full Documentation

> A comprehensive catalog of text-to-speech voices across multiple providers, languages, and accents. Curated for builders and engineers working on the cutting edge of TTS technology with 4,456+ voices spanning 92 languages and 264 regional dialects across 115+ countries.

TTS Library provides access to thousands of high-quality voice samples from leading TTS providers (Telnyx, MiniMax, Inworld, Azure, Rime, Resemble, AWS, FishAudio, XAI, and Humain). Each voice includes metadata for provider, language, country, accent/dialect, gender, and direct audio previews. The platform enables efficient discovery and evaluation of voices for TTS implementation and research across global markets.

Live at: https://ttslibrary.com

## Language Coverage

**92 total languages** | **264 distinct accents/dialects** | **115+ countries**

**European Languages:**
- Germanic: German (Standard, Austrian, Swiss), Dutch (Flemish), English (20+ regional variations), Swedish, Norwegian, Danish, Icelandic
- Romance: Spanish (Spain, Mexican, Argentinian, Colombian, Venezuelan, etc.), French (France, Canadian, Belgian, Swiss), Portuguese (Brazilian, European), Italian, Romanian, Catalan, Galician
- Slavic: Russian, Polish, Czech, Slovak, Croatian, Serbian, Bulgarian, Ukrainian, Slovenian, Bosnian, Macedonian
- Other: Greek, Hungarian, Finnish, Estonian, Latvian, Lithuanian, Albanian, Welsh, Irish, Maltese, Basque

**Asian Languages:**
- Chinese: Mandarin (Mainland, Taiwan, Guangxi, Henan, Liaoning, Shaanxi, Shandong), Cantonese, Wu Chinese (Shanghainese), Sichuanese
- East Asian: Japanese, Korean
- Southeast Asian: Vietnamese, Thai, Indonesian, Malaysian, Javanese, Sundanese, Filipino, Khmer, Lao, Burmese
- South Asian: Hindi, Bengali, Tamil (India, Malaysia, Singapore, Sri Lanka), Telugu, Marathi, Gujarati, Kannada, Malayalam, Punjabi, Urdu (India, Pakistan), Nepali, Sinhala, Assamese, Odia

**Middle Eastern & African Languages:**
- Arabic: Standard Arabic plus 15+ regional dialects (Egyptian, Saudi, Moroccan, Lebanese, Iraqi, Emirati, Bahraini, Kuwaiti, Omani, Qatari, Algerian, Libyan, Tunisian, Syrian, Yemeni)
- Other Middle Eastern: Hebrew, Persian, Turkish, Armenian, Georgian, Azerbaijani, Kurdish (Pashto)
- African: Swahili (Kenya, Tanzania), Zulu, Afrikaans, Amharic, Somali, Hausa

**Other Global Languages:**
- Mongolian, Kazakh, Uzbek, Inuktitut (Latin/Syllabics), various minority and indigenous languages

## Providers

| Provider | Voices | Models/Tiers |
|----------|--------|--------------|
| **Telnyx** | 1,221 | Ultra, Natural, NaturalHD, KokoroTTS, Qwen3TTS, LibriTTS, Bayan, Sukhan |
| **MiniMax** | 996 | speech-02-turbo, speech-2.6-turbo, speech-2.8-turbo |
| **Inworld** | 807 | TTS2, Max, Mini |
| **Azure** | 769 | Neural, Multilingual, DragonHD (Latest/Flash/Omni), MAI-Voice-1 |
| **Rime** | 301 | ArcanaV3, Coda |
| **Resemble** | 198 | Pro, Turbo |
| **AWS** | 123 | Polly Neural, Polly Standard |
| **FishAudio** | 30 | s1, s2-pro, s2.1-pro |
| **Humain** | 6 | — |
| **XAI** | 5 | rex, ara, leo, sal, eve |

Note: Rime is its own provider (`provider: rime`), distinct from Telnyx, with its own filter in the UI.

## Voice Data Schema

Each voice object in the catalog contains:

| Field | Type | Present | Description | Example |
|-------|------|---------|-------------|---------|
| id | string | Always | Unique voice identifier | `AWS.Polly.Joanna-Neural` |
| name | string | Always | Human-readable voice name | `Joanna` |
| language | string | Always | BCP-47 language-region code | `en-US` |
| provider | string | Always | TTS provider | `aws`, `azure`, `minimax`, `telnyx`, `inworld`, `rime`, `resemble`, `fishaudio`, `xai`, `humain` |
| gender | string | Always | Voice gender | `Female`, `Male`, `Neutral` |
| model_id | string | Telnyx, MiniMax | Model tier name | `Ultra`, `speech-2.6-turbo` |
| age | string | MiniMax only | Voice age category | `adult` |
| label | string | MiniMax only | Provider description of the voice | `An expressive adult male voice...` |
| accent | string | MiniMax raw, all via API | Accent classification | `American`, `British` |

### Enriched Fields (added by API at request time)

These fields are not in the raw data but are added by `/api/voices`:

| Field | Description |
|-------|-------------|
| language_name | Resolved language name from locale mapping (e.g., `English`) |
| country_name | Resolved country name from locale mapping (e.g., `United States`) |
| accent | Accent classification from voice_accents.json (e.g., `American`) |
| description | Short description, 90 char limit (e.g., `AWS Neural American English female voice`) |
| uniqueKey | Unique key for rendering (same as id) |

### Voice ID Conventions

- **AWS:** `AWS.Polly.{Name}-Neural` or `AWS.Polly.{Name}` (standard)
- **Azure:** `Azure.{locale}-{Name}Neural`
- **MiniMax:** `Minimax.{model}.{VoiceName}`
- **Telnyx:** `Telnyx.{Model}.{name}`
- **Inworld:** `Inworld.{Model}.{name}`
- **Rime:** `Rime.{Model}.{name}` (its own `provider: rime`, not Telnyx)
- **Resemble:** `Resemble.{Model}.{name}`
- **FishAudio:** `FishAudio.{Model}.{name}`
- **XAI:** `XAI.{name}`

## API Reference

Base URL: `https://ttslibrary.com`

### GET /api/voices

Filters, deduplicates, and paginates the voice catalog. Voices are sorted with Telnyx Ultra first.

**Query Parameters:**

| Parameter | Type | Default | Description |
|-----------|------|---------|-------------|
| search | string | `""` | Full-text search across name, ID, language, provider, accent |
| providers | string | `""` | Comma-separated: `telnyx,aws,azure,minimax,inworld,rime,resemble,fishaudio,xai,humain` |
| genders | string | `""` | Comma-separated: `Female,Male,Neutral` |
| language | string | null | Language name filter: `English`, `Spanish`, `Chinese`, etc. |
| country | string | null | Country name filter: `United States`, `Spain`, `Mexico`, etc. |
| model | string | null | Model/tier name: `Ultra`, `Neural`, `speech-2.6-turbo`, etc. |
| accent | string | null | Accent filter: `American`, `British`, `Castilian`, etc. |
| ages | string | `""` | Comma-separated age filter: `adult` |
| limit | int | 48 | Results per page |
| offset | int | 0 | Pagination offset |

**Response:**

```json
{
  "voices": [
    {
      "id": "Telnyx.Ultra.Bella",
      "name": "Bella",
      "language": "en-US",
      "provider": "telnyx",
      "model_id": "Ultra",
      "gender": "Female",
      "uniqueKey": "Telnyx.Ultra.Bella",
      "accent": "American",
      "description": "Telnyx Ultra American English female voice",
      "language_name": "English",
      "country_name": "United States"
    }
  ],
  "total": 4456,
  "limit": 48,
  "offset": 0,
  "hasMore": true,
  "performance": {
    "total_voices": 4456,
    "filtered_count": 4456,
    "returned_count": 48
  }
}
```

**Example Queries:**

```
# All Spanish voices
GET /api/voices?language=Spanish

# Female Telnyx Ultra voices
GET /api/voices?providers=telnyx&genders=Female&model=Ultra

# Search by name
GET /api/voices?search=joanna

# British accent English voices
GET /api/voices?language=English&accent=British

# All voices (no pagination)
GET /api/voices?limit=3000
```

### GET /api/metadata

Returns all available filter values and statistics. Values are derived dynamically from the voice catalog.

**Response:**

```json
{
  "total_voices": 4456,
  "providers": ["aws", "azure", "fishaudio", "humain", "inworld", "minimax", "resemble", "rime", "telnyx", "xai"],
  "genders": ["Female", "Male", "Neutral"],
  "languages": ["Afrikaans", "Albanian", "Amharic", "..."],
  "countries": ["Afghanistan", "Algeria", "..."],
  "accents": ["Algerian Arabic", "American", "Argentine", "..."],
  "stats": {
    "total_voices": 4456,
    "total_languages": 92,
    "total_countries": 115,
    "total_providers": 10,
    "total_accents": 264
  }
}
```

### GET /api/audio-url

Returns an authenticated presigned URL for voice audio preview from Telnyx Cloud Storage.

**Query Parameters:**

| Parameter | Type | Required | Description |
|-----------|------|----------|-------------|
| voiceId | string | Yes | Voice ID from catalog (e.g., `AWS.Polly.Joanna-Neural`) |

**Response:**

```json
{
  "url": "https://us-central-1.telnyxcloudstorage.com/...",
  "voiceId": "AWS.Polly.Joanna-Neural",
  "filename": "AWS_Polly_Joanna-Neural.mp3",
  "source": "presigned"
}
```

## Core Pages

- [Main Voice Directory](https://ttslibrary.com/) - Browse and search all available TTS voices with filters
- [Terms of Service](https://ttslibrary.com/terms) - Usage terms and conditions
- [Privacy Policy](https://ttslibrary.com/privacy) - Privacy policy and data handling

## Featured Language Collections

| Language | URL | Description |
|----------|-----|-------------|
| Spanish | https://ttslibrary.com/spanish | 10+ accent variations: Spain, Mexico, Argentina, Colombia, Venezuela |
| Portuguese | https://ttslibrary.com/portuguese | Brazilian Portuguese and European Portuguese variants |
| Chinese | https://ttslibrary.com/chinese | Mandarin regional variations, Cantonese, Wu Chinese |
| Korean | https://ttslibrary.com/korean | Korean language TTS voices |
| French | https://ttslibrary.com/french | France, Canada, Belgium, and Switzerland accents |
| Japanese | https://ttslibrary.com/japanese | Japanese language TTS voices |
| German | https://ttslibrary.com/german | Standard German, Austrian, and Swiss German |
| Italian | https://ttslibrary.com/italian | Italian language TTS voices |
| Russian | https://ttslibrary.com/russian | Russian language TTS voices |
| Indonesian | https://ttslibrary.com/indonesian | Indonesian, Malaysian, Javanese, Sundanese |

## Agent Access

### Content Negotiation

Request any page with the `Accept: text/markdown` header to receive markdown instead of HTML:

```bash
# HTML response (default)
curl https://ttslibrary.com/

# Markdown response
curl -H "Accept: text/markdown" https://ttslibrary.com/

# Markdown for a language page
curl -H "Accept: text/markdown" https://ttslibrary.com/spanish
```

### Direct Markdown Endpoints

| Endpoint | Description |
|----------|-------------|
| /api/md/sitemap | Markdown table of contents for all pages |
| /api/md | Markdown overview of the site |
| /api/md/voices | Markdown voice directory with stats |
| /api/md/language?lang=spanish | Markdown for a specific language page |

### Documentation Files

| File | Description |
|------|-------------|
| /llms.txt | Concise site overview |
| /llms-full.txt | Full documentation (this file) |

## Technical Resources

- [XML Sitemap](https://ttslibrary.com/sitemap.xml) - Complete site structure for search engines
- [Robots.txt](https://ttslibrary.com/robots.txt) - Crawling permissions and sitemap reference
- [LLMs.txt](https://ttslibrary.com/llms.txt) - Concise overview for AI agents
- [LLMs Full](https://ttslibrary.com/llms-full.txt) - This file