Voices and languages

SignalWire integrates natively with leading third-party text-to-speech (TTS) providers. This guide describes supported engines, voices, and languages. Refer to each provider’s documentation for up-to-date model details and service information.

Browse and audition voices

Choose a provider to browse and audition its full voice catalog. Press play to audition a voice, and use copy config to grab the engine and voice values for your SWML or SDK code. Each provider’s complete voice list lives on its reference page, linked in the table below.

Only voices from SignalWire’s catalog resolve on the platform — voice IDs from your own provider account will not work. See Custom and cloned voices for supported options.

Compare providers and models

SignalWire’s TTS providers offer a wide range of voice engines optimized for various applications. Select a provider, model, and voice according to the following considerations:

Language support: At time of writing, engine language support is as follows. Consult each provider’s reference documentation for the most up-to-date information.

Rime voices support English, Spanish, French, German, Portuguese, Hindi, Arabic, and Japanese.
Deepgram voices support English, Spanish, German, French, Dutch, Italian, and Japanese.
Amazon Polly, Azure, Cartesia, and Google Cloud offer a wide range of supported languages.
Inworld voices support English plus Arabic, Chinese (Mandarin), Dutch, French, German, Hebrew, Hindi, Italian, Japanese, Korean, Polish, Portuguese, Russian, and Spanish.
MiniMax voices span more than 20 languages, including English, Spanish, Portuguese, French, German, Italian, Chinese, Japanese, Korean, and many more, with automatic language detection.
Fish Audio voices cover roughly 83 languages on the s2.1-pro model, with automatic language detection.
Mistral voices support English, French, Spanish, Portuguese, Italian, Dutch, German, Hindi, and Arabic.
Smallest.ai voices support English, Spanish, Hindi, and nine other Indian languages.
Speechify voices are multilingual on the simba-3.0 model.
Groq-hosted Orpheus voices support English and Arabic (Saudi), one language per model.
All ElevenLabs, OpenAI, and Grok voices are fully multilingual.

SSML support: Google Cloud and Amazon Polly support SSML (Speech Synthesis Markup Language) as a string wrapped in <speak> tags. Consult Google Cloud’s SSML docs for details. Refer to the Amazon Polly docs for more information on using SSML and supported SSML tags.

Use voice identifier strings

Compose voice identifier strings using the following general format:

engine.voice:model

Identifier	Description
`engine` required	The TTS provider (e.g., `elevenlabs`, `rime`, `openai`)
`voice` required	The voice identifier (name or ID depending on engine)
`model` optional	Model variant (not all engines support this)

Since voice ID strings are case insensitive, the following strings are equivalent:

gcloud.en-US-Neural2-A
gcloud.en-us-neural2-a
GCLOUD.EN-US-NEURAL2-A

For detailed instructions for each provider, consult the voice ID references linked in the Usage column of the below table.

TTS provider	Sample voice ID string	Usage
Amazon Polly	`amazon.Joanna-Neural`	Reference
Azure	`azure.en-US-AvaNeural`	Reference
Cartesia	`cartesia.a167e0f3-df7e-4d52-a9c3-f949145efdab`	Reference
Deepgram	`deepgram.aura-asteria-en`	Reference
ElevenLabs	`elevenlabs.thomas`	Reference
Fish Audio	`fish.802e3bc2b27e49c2995d23ef70e6ac89:s2.1-pro`	Reference
Google Cloud	`gcloud.en-US-Casual-K`	Reference
Grok	`grok.eve`	Reference
Groq	`groq-tts.autumn`	Reference
Inworld	`inworld.Lauren:inworld-tts-1.5-mini`	Reference
MiniMax	`minimax.English_CalmWoman:speech-2.6-turbo`	Reference
Mistral	`mistral.<voiceId>`	Reference
OpenAI	`openai.alloy`	Reference
Rime	`rime.luna:coda`	Reference
Smallest.ai	`smallest.sophia:lightning_v3.1`	Reference
Speechify	`speechify.alicia:simba-3.0`	Reference

Custom and cloned voices

SignalWire’s TTS integrations run on SignalWire’s own accounts with each provider. When you set a voice string like elevenlabs.<voice_id> or cartesia.<voice_id>:<model>, SignalWire resolves the voice ID against its own provider account — not yours.

Only voices from SignalWire’s catalog resolve on the platform. A voice ID from your own provider account — including custom or cloned voices — is not visible to SignalWire’s account. It will not resolve, and TTS falls back to a default voice at call time. This applies even though a cloned-voice ID looks identical in format to a catalog voice ID.

ElevenLabs custom and cloned voices can be imported into SignalWire’s catalog by request. Contact Support to request a voice import. Once the voice is imported, use it with the standard voice string format.

Pricing

See the Voice API Pricing page for up-to-date pricing information.