Lightning
Our fastest model, optimized for low-latency applications. It can generate 10 seconds of audio in just 100 milliseconds, making it ideal for real-time applications such as voicebots and interactive systems.
Lightning v2
An upgrade from the Lightning Large model, offering improved performance and quality. It supports 16 languages, making it suitable for a wider range of applications requiring expressive and high-quality speech synthesis.
Lightning Large [⚠️ To be Deprecated]
Offers more emotional depth and expressiveness compared to the Lightning model. It supports voice cloning and has a latency of just under 300 milliseconds, making it suitable for applications requiring high-quality, expressive speech.
Model Overview
| Model ID | Description | Languages Supported |
|---|---|---|
| lightning | Fastest model with an RTF of 0.01, generating 10 seconds of audio in 100 ms. | English Hindi |
| lightning-large | More emotional depth and expressiveness, supports voice cloning, latency under 300 ms. | English Hindi |
| lightning-v2 | Supports 16 languages, currently in beta. | English Hindi Tamil Kannada Gujarati Bengali Marathi German French Spanish Italian Polish Dutch Russian Arabic Hebrew |
Note: The API uses ISO 639-1 language codes - Set 1 (2-letter codes) to specify supported languages.