> ## Documentation Index
> Fetch the complete documentation index at: https://api-docs.ollang.com/llms.txt
> Use this file to discover all available pages before exploring further.

# ElevenLabs V4 Text-to-Speech API

> ElevenLabs Eleven v4 — the default voice engine in Ollang, with 90+ languages, directed delivery through audio tags, and the same cloned voices as Multilingual V2.

**Eleven v4** is the model ElevenLabs describes as its most emotive, highest-quality speech synthesis. It is the default voice engine for AI Dubbing in Ollang, and it extends the ElevenLabs voice library to **90+ languages** with far more control over how each line is performed.

> [!NOTE]
>
> **Available in Ollang Workflows** as `elevenlabsV4` — AI Dubbing · **Default** voice engine. See the [Text-to-Speech catalog](https://api-docs.ollang.com/providers/text-to-speech.md).

## Model and coverage

| | |
| - | - |
| **Model** | Eleven v4 (`eleven_v4`) |
| **Languages** | 90+ |
| **Character limit** | 10,000 per request |
| **Voices** | Your ElevenLabs cloned voices plus the ElevenLabs premade library |
| **Low-latency variant** | Eleven v4 Turbo (`eleven_v4_turbo`), not used by Ollang's `elevenlabsV4` provider |

ElevenLabs reports the largest quality gains over earlier models in Japanese, Brazilian Portuguese, Mandarin, and Cantonese. Ollang only offers V4 voices for target languages the model supports.

## Key capabilities

* **Audio tags and delivery directions.** Inline tags such as `[laughs]` or `[phone buzzing]`, and free-form directions such as `[sleepy, drowsy voice]`, steer emotion, pacing, and non-verbal sounds. ElevenLabs reports that V4 follows these more accurately than earlier models.
* **Shared voice library.** Voice IDs are model-independent on the ElevenLabs side, so voices cloned for [Multilingual V2](https://api-docs.ollang.com/apis/tts-apis/elevenlabs.md) work on V4 without being recreated.
* **Instant and professional voice clones** for carrying a character or brand voice across every target language.
* **Multi-speaker dialogue** and improved phoneme (IPA) support for names and technical terms.

## Where it fits in a workflow

V4 is the default because it combines ElevenLabs' voice quality with language coverage close to the widest engines in the catalog. It suits drama, entertainment, marketing, and character-driven content, and it is the natural upgrade for any Workflow currently running Multilingual V2: the voices carry over and the language list roughly triples.

For the widest language reach, [Gemini TTS](https://api-docs.ollang.com/apis/tts-apis/gemini-2-5-tts.md) covers 99 languages. For performance driven entirely by written direction, compare it with [Fish Audio](https://api-docs.ollang.com/apis/tts-apis/fish-audio.md).

## Trade-offs

* **Newer model line**, with less accumulated production history than Multilingual V2.
* **Direction is where the extra quality comes from.** Undirected lines are natural but closer to the field, so budget time for delivery cues on performance-heavy content.

## Reference

* [Text-to-Speech provider catalog](https://api-docs.ollang.com/providers/text-to-speech.md)
* [ElevenLabs Multilingual V2](https://api-docs.ollang.com/apis/tts-apis/elevenlabs.md)
* [ElevenLabs models](https://elevenlabs.io/docs/overview/models)
* [Introducing Eleven v4](https://elevenlabs.io/blog/eleven-v4)


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