Skip to content



Imagine giving your chatbot or web app the ability to talk – no mics, no actors – just crisp, clear audio generated from text. LMNT provides a text-to-speech (TTS) API and high-quality voices to make your app smoother and more engaging with human-like intonation and delivery.

You can also stream text to our servers and receive synthesized audio in real-time for your users at low latency (<300ms) with our WebSocket API.

Give each bot, character, or app a personality with a unique voice. We support a wide library of voices and the ability to instantly clone your own from 30 secs of audio (or more for a professional clone).

Try it at

Additional Information

This integration will set up your Vercel project(s) with an LMNT API key for your application to begin accessing our text-to-speech via REST or WebSockets APIs. You can also install our Node or Python SDKs:

A typical request involves choosing a voice, sending text, and then receiving audio to use in your application.

After you have completed the integration setup, you’ll see an environment variable named LMNT_API_KEY in your project's settings page.

For more information, visit our Vercel LMNT integration docs.

For instant help or questions, ping us on Discord or We’re adding more languages as well – get in touch for early access and to request a specific language.

Get started with Vercel and LMNT.