ElevenLabs TTS

Yes, Eleven v3 is the underlying expressive speech synthesis architecture powering this node. It introduces dynamic emotional prosody, natural micro-breaths, and inline theatrical tags like [whispers] and [laughs] to produce lifelike vocal performances.

Set stability between 0.30 and 0.50 for dramatic character acting, dialogue, and scripts using inline emotion tags. For e-learning, news reading, technical explainers, or any project requiring uniform, repeatable delivery across lines, set stability to between 0.70 and 0.85.

Place natural language tags in brackets directly before the dialogue—such as [whispers], [sighs], [laughs], or [excited]. Combine these tags with punctuation like em dashes (—) for pauses and keep stability around 0.30–0.50 so the model can freely interpret the dramatic cues.

Pass a two-letter ISO 639-1 code such as en, es, fr, de, or ja into the language_code parameter. This enforces regional pronunciation rules while preserving the acoustic persona and timbre of your chosen voice.

Avoid this model when sub-100ms turn latency is strictly required, such as in real-time telephony agents. It is also not recommended for unstitched multi-hour continuous audiobooks where seamless prosodic uniformity across separate API chunks without voice drift is critical.