> ## Documentation Index
> Fetch the complete documentation index at: https://supertone-b89202c8.mintlify.site/llms.txt
> Use this file to discover all available pages before exploring further.

# Release notes

> Latest updates and changes to the Supertone API.

**2026-06-16**

## Command-line interface and MCP server

Two new ways to use Supertone TTS beyond the SDKs: a terminal CLI and an MCP server for AI agents. Both wrap the same Text-to-Speech API.

### What's new

1. **Supertone CLI**
   * Synthesize from strings, files, or stdin; stream to your speakers; batch whole directories; and manage voices, usage, and credits — all from the terminal.
   * `pip install supertone-cli`
   * See [CLI](/en/docs/developer-tools/cli).

2. **Supertone MCP server**
   * Exposes the TTS API as composable [Model Context Protocol](https://modelcontextprotocol.io) tools, so AI agents in Claude, Cursor, Claude Code, VS Code, and Windsurf can discover voices, estimate cost, clone voices, and synthesize speech.
   * `uvx supertone-mcp`
   * See [MCP](/en/docs/developer-tools/mcp).

**2026-05-19**

## New TTS model: `supertonic_api_3`

The Supertone API now supports `supertonic_api_3`, the next-generation successor to `supertonic_api_1`. It keeps the ultra-low latency profile of its predecessor while dramatically improving speech stability and expanding language coverage.

### What's new

1. **`supertonic_api_3` — Successor to `supertonic_api_1`**
   * Inherits the ultra-low latency profile of `supertonic_api_1`
   * **Significantly improved speech stability** — far fewer pronunciation and reading errors
   * Trained differently from the open-weights Supertonic 3 release, so this is a distinct model
   * Supports **31 languages**: `en`, `ko`, `ja`, `ar`, `bg`, `cs`, `da`, `de`, `el`, `es`, `et`, `fi`, `fr`, `hi`, `hr`, `hu`, `id`, `it`, `lt`, `lv`, `nl`, `pl`, `pt`, `ro`, `ru`, `sk`, `sl`, `sv`, `tr`, `uk`, `vi`
   * Voice settings: `speed` only (same as `supertonic_api_1`)

2. **`supertonic_api_1` is now the legacy supertonic model**
   * Still available for existing integrations
   * New projects should use `supertonic_api_3`

3. **Model selection expanded** — TTS requests now accept `sona_speech_2`, `sona_speech_2_flash`, `supertonic_api_3`, `supertonic_api_1`, and `sona_speech_1`.

See [Models](/en/docs/core-concepts/models) for full comparison and selection guidance.

**2025-12-08**

## New TTS models: `sona_speech_2`, `sona_speech_2_flash`, and `supertonic_api_1`

The Supertone API now supports three new speech synthesis models, expanding multilingual capabilities and offering more flexible generation options.

### What's new

1. **`sona_speech_2` — Multilingual flagship**
   * Supports **23 languages**
   * Improved speech quality and stability
   * Supports `include_phonemes` and `normalized_text`
   * Voice settings: all parameters except `subharmonic_amplitude_control`

2. **`sona_speech_2_flash` — Low-latency variant**
   * Same 23 languages as `sona_speech_2`
   * Optimized for lower latency
   * Voice settings: `pitch_shift`, `pitch_variance`, `speed`, `duration`

3. **`supertonic_api_1` (beta) — Lightweight low-latency model**
   * Supports `en`, `ko`, `ja`, `es`, `pt`
   * Designed for fastest inference and simplest setup
   * Voice settings: `speed` only

4. **Model selection expanded** — TTS requests now accept `sona_speech_2`, `sona_speech_2_flash`, `supertonic_api_1`, and `sona_speech_1`.

See [Models](/en/docs/core-concepts/models) for full comparison and selection guidance.

**2025-07-31**

## Custom voices and streaming TTS (beta)

Several new features have been added to the Supertone API. In this release, **Preset Voices** and **Custom Voices** have been separated, and voice cloning via API is now supported. A beta version of streaming TTS has also been released.

### What's new

1. **Preset vs custom voice separation**
   * Preset voices are designed and provided by Supertone.
   * Custom voices are clones registered and managed by the customer.
   * New endpoints under `/v1/custom-voices/...` for managing customer-owned voices.

2. **Voice cloning via the API**
   * Voice cloning, previously only available in Supertone Play, is now supported via API.
   * See [Create cloned voice](/en/api-reference/endpoints/create-cloned-voice).

3. **Streaming TTS (beta)**
   * A beta version of the streaming TTS API is available, ideal for real-time applications such as AI chatbots and character-based chat.
   * See [Stream speech](/en/docs/text-to-speech/stream-speech).

**2025-05-28**

## Supertone API grand open

The Supertone API has officially launched after completing its closed beta phase. This release features a complete overhaul of the service structure and API usability, evolving into a more intuitive and scalable voice generation platform through integration with Play and unification of the voice system.

![](https://mintcdn.com/supertone-b89202c8/0gAKILFZFHulmmA8/images/supertone-api-grand-open.png?fit=max\&auto=format\&n=0gAKILFZFHulmmA8\&q=85\&s=1a8c108b3cc1569bd3b87aacc44fc0b1)

### Major changes

1. **Voice ID structure redesign**
   * Changed from `{character-language-style}` combination → unified to **character-based `voice_id`**.
   * `language` and `style` parameters moved into the request body for flexible calling.

2. **New TTS model**
   * Replaced `pro` and `turbo` → unified into a single `sona_speech_1` model.
   * Improved speech quality; text length limit increased from 200 to **300 characters**.

3. **Cloned voice API support**
   * Cloned voices created in Play can now be called via the API.
   * Automatic integration within the same account, with permission-based security.

4. **Play and pricing integration**
   * API usage applies Play's credit system directly.
   * Shared credits and unified billing — no separate pricing plans.

5. **Enhanced voice search**
   * Filtering and exploration via `GET /voices` and `GET /voices/search`.
   * Each voice exposes supported languages, styles, and samples.

6. **Predict Duration API**
   * Estimate speech length before generation — no credits deducted.

7. **New user guide and documentation**
   * Practical documentation including Quickstart, Voice Selection, Error Handling, and FAQ.
   * Complete documentation renewal focused on the Play ↔ API integration flow.

**2025-01-23**

## Voice search update

New search features make it easier to find the voices you want.

### What's new

* New endpoint for searching voices with filters and keywords.
* Filter by language, gender, age, and use case.
* Keyword search across voice names and descriptions.
* Page-based pagination for large result sets.

See [Search voices](/en/api-reference/endpoints/search-voices).

**2024-12-23**

![](https://mintcdn.com/supertone-b89202c8/0gAKILFZFHulmmA8/images/supertone-api-landing-page.png?fit=max\&auto=format\&n=0gAKILFZFHulmmA8\&q=85\&s=1e13e66b7e2d50b36c928603b7986db5)

## Supertone API closed beta launch

Welcome to the **Supertone API** closed beta — Supertone's voice AI technology as a service.

### Key features

1. Text-to-speech voice synthesis service.
2. Voice list inquiry and usage check APIs.

### Endpoints

* [List voices](/en/api-reference/endpoints/list-voices) — retrieve available voices.
* [Create speech](/en/api-reference/endpoints/text-to-speech) — convert text to speech.
* [Get voice usage](/en/api-reference/endpoints/get-voice-usage) — check API usage.

### Getting started

1. Sign up at [console.supertoneapi.com](https://console.supertoneapi.com).
2. Apply for closed beta participation.
3. Issue an API key from the console.
4. Integrate and use the API.

### Console features

* API key issuance and management.
* Minute-based API usage monitoring.

For more details, refer to the [docs](/en/docs/overview).


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.