> For a complete page index, fetch https://docs.synthflow.ai/llms.txt. For full documentation content, fetch https://docs.synthflow.ai/llms-full.txt.

## ElevenLabs V3 - Audio Tags

Synthflow now supports **ElevenLabs V3** as a voice model, bringing inline audio tag control to your agents.

### What's new

* **Global audio tags** — Set a persistent voice character for the entire call, such as `[American accent]` or `[warm and friendly tone]`, via the new **Global Audio Tags** field in Voice Configuration or the `global_audio_tags` API parameter.
* **Situational audio tags** — Place tags like `[excited]`, `[whispers]`, `[laughs]`, and `[sighs]` inline in your prompts to control delivery at specific moments. Each tag affects approximately 4-5 words before returning to normal delivery.

### How to use it

Select **ElevenLabs V3** as the voice model in [Voice Configuration](/voice-configuration) and start using audio tags in your prompts. See the full guide: [Audio Tags (ElevenLabs V3)](/voice-configuration#audio-tags).

### API changes

* `voice_synthesizer` now accepts `eleven_v3` as a valid value.
* New optional `global_audio_tags` field on the Agent object for setting persistent voice tags via the API.