The voice collection

Find a voice
that feels like you.

Hear the voices our speech service can speak, each reading the same customer line. Find a direction for your brand, then bring the voice, knowledge, and actions together in your customer workflow.

Explore in Voice Studio Explore languages

Listen, compare, then build

A few voices. Different possibilities.

We generated each preview through our own provider accounts and serve it from this site. A complete Voice workflow also includes recognition, reasoning, tools, and turn-taking.

Search voice previews

Language

10 voice previews

Alloy

OpenAI · tts-1

English

Availability: Default voice The voice our live speech service is configured to use.
API voice: alloy

Start here: this is the voice live conversations use today. Listen to whether the two appointment options stay easy to tell apart.

AI-generated voiceReads: “Hi, thanks for calling. I can help you move your appointment. Would you prefer Tuesday morning or Thursday afternoon?”

Sarah

ElevenLabs · eleven_flash_v2_5

English · American accent

Availability: Configured The voice set for this provider in our live speech service.
Described by ElevenLabs as: Mature, reassuring, confident
API voice: EXAVITQu4vr4xnSDxMaL

The ElevenLabs voice our speech service falls back to. Compare pacing and sentence endings with Alloy on the same line.

AI-generated voiceReads: “Hi, thanks for calling. I can help you move your appointment. Would you prefer Tuesday morning or Thursday afternoon?”

Neural2-D

Google Cloud Text-to-Speech · Neural2

English · US

Availability: Configured The voice set for this provider in our live speech service.
API voice: en-US-Neural2-D

Listen to emphasis on the times and the pause before the question. Shortlist against your own service tone.

AI-generated voiceReads: “Hi, thanks for calling. I can help you move your appointment. Would you prefer Tuesday morning or Thursday afternoon?”

Echo

OpenAI · tts-1

English

Availability: Supported by the integration Our integration accepts this voice. It is not yet selectable per workflow.
API voice: echo

Listen to the rhythm across the whole line and how the final question lands.

AI-generated voiceReads: “Hi, thanks for calling. I can help you move your appointment. Would you prefer Tuesday morning or Thursday afternoon?”

Fable

OpenAI · tts-1

English

Availability: Supported by the integration Our integration accepts this voice. It is not yet selectable per workflow.
API voice: fable

Pay attention to emphasis and transitions between phrases.

AI-generated voiceReads: “Hi, thanks for calling. I can help you move your appointment. Would you prefer Tuesday morning or Thursday afternoon?”

Onyx

OpenAI · tts-1

English

Availability: Supported by the integration Our integration accepts this voice. It is not yet selectable per workflow.
API voice: onyx

Compare the weight of this voice with the others on the same appointment line.

AI-generated voiceReads: “Hi, thanks for calling. I can help you move your appointment. Would you prefer Tuesday morning or Thursday afternoon?”

Nova

OpenAI · tts-1

English

Availability: Supported by the integration Our integration accepts this voice. It is not yet selectable per workflow.
API voice: nova

Compare clarity on the days and times, the details a caller must not mishear.

AI-generated voiceReads: “Hi, thanks for calling. I can help you move your appointment. Would you prefer Tuesday morning or Thursday afternoon?”

Shimmer

OpenAI · tts-1

English

Availability: Supported by the integration Our integration accepts this voice. It is not yet selectable per workflow.
API voice: shimmer

Listen to pacing and consider how much detail your customers need to hear.

AI-generated voiceReads: “Hi, thanks for calling. I can help you move your appointment. Would you prefer Tuesday morning or Thursday afternoon?”

Sarah

ElevenLabs · eleven_flash_v2_5

Hindi · हिन्दी

Availability: Configured The voice set for this provider in our live speech service.
Described by ElevenLabs as: Mature, reassuring, confident
API voice: EXAVITQu4vr4xnSDxMaL

The configured ElevenLabs voice reading Hindi through a multilingual model. ElevenLabs labels Sarah an American English voice, so judge pronunciation with your own names and places.

AI-generated voiceReads: “नमस्ते, कॉल करने के लिए धन्यवाद। क्या आप मंगलवार सुबह आना चाहेंगे या गुरुवार दोपहर?”

Nova

OpenAI · tts-1

Hindi · हिन्दी

Availability: Supported by the integration Our integration accepts this voice. It is not yet selectable per workflow.
API voice: nova

The same Hindi line from an OpenAI voice. Compare pronunciation and the space between phrases with Sarah.

What you can rely on today

Language coverage and availability.

Every preview reads a short appointment line, so you compare voices on the same words. The Hindi line means: “Hello, thank you for calling. Would you like to come on Tuesday morning or Thursday afternoon?”

English

Previews from ElevenLabs, OpenAI, and Google Cloud Text-to-Speech. Alloy is the default voice for live conversations; Sarah and Neural2-D are the voices configured for the other two providers.

Hindi

Sarah and Nova reading the same Hindi line through multilingual models. We do not yet offer a dedicated Hindi voice, so test your own names, places, and numbers before you launch.

Choosing per workflow

Voices marked “Supported by the integration” are accepted by our provider integration but are not yet selectable per workflow. Contact us if one of them is the right fit.

Self-hosted voices you can add

Our speech service includes integrations for self-hosted engines. None is running today, and each needs its own model server before it can speak.

Piper · Self-hosted, needs a model server before it can speak.
Kokoro · Self-hosted, needs a model server before it can speak.
XTTS · Self-hosted, needs a model server before it can speak.