ElevenLabs

Use ElevenLabs text-to-speech voices in Vapi.

ElevenLabs provides text-to-speech for Vapi voice agents.

You can use ElevenLabs through Vapi’s default integration, or connect your own account in Integrations.

Supported capability

CapabilityProvider value
Text-to-speech11labs

Text-to-speech

Set voice.provider to 11labs, voice.model to a supported model, and voice.voiceId to a voice. ElevenLabs offers many voices and supports voice cloning. You can browse and copy a voice’s ID from the Voice Library; the default voice is sarah.

$curl -X PATCH "https://api.vapi.ai/assistant/ASSISTANT_ID" \
> -H "Authorization: Bearer $VAPI_API_KEY" \
> -H "Content-Type: application/json" \
> -d '{
> "voice": {
> "provider": "11labs",
> "model": "eleven_flash_v2_5",
> "voiceId": "sarah"
> }
> }'

For additional API configuration options, review the ElevenLabsVoice fields in the Create Assistant API reference.

Supported models

ModelModel ID
Eleven Flash v2eleven_flash_v2
Eleven Flash v2.5eleven_flash_v2_5
Eleven Multilingual v2eleven_multilingual_v2
Eleven v3eleven_v3

Configuration options

OptionAccepted valuesDescription
stability0 to 1Controls consistency between generations.
similarityBoost0 to 1Controls how closely the output matches the selected voice.
style0 to 1Controls style exaggeration.
useSpeakerBoosttrue or falseEnables speaker similarity enhancement.
speed0.7 to 1.2Controls speaking speed.
optimizeStreamingLatency0 to 4Controls streaming-latency optimization. The default is 3.
enableSsmlParsingtrue or falseEnables supported SSML pronunciation syntax. The default is false.
autoModetrue or falseEnables automatic mode selection. The default is false.

Supported languages

ElevenLabs models infer the language from the text. Language coverage depends on the selected model.

The following language coverage is published for models supported by Vapi. See ElevenLabs models for provider details.

LanguageLanguage codeEleven Flash v2Eleven Flash v2.5Eleven Multilingual v2Eleven v3
AfrikaansafX
ArabicarXXX
ArmenianhyX
AssameseasX
AzerbaijaniazX
BelarusianbeX
BengalibnX
BosnianbsX
BulgarianbgXXX
CatalancaX
CebuanoX
ChichewanyX
CroatianhrXXX
CzechcsXXX
DanishdaXXX
DutchnlXXX
EnglishenXXXX
EstonianetX
FilipinoXXX
FinnishfiXXX
FrenchfrXXX
GalicianglX
GeorgiankaX
GermandeXXX
GreekelXXX
GujaratiguX
HausahaX
HebrewheX
HindihiXXX
HungarianhuXX
IcelandicisX
IndonesianidXXX
IrishgaX
ItalianitXXX
JapanesejaXXX
JavanesejvX
KannadaknX
KazakhkkX
KirghizkyX
KoreankoXXX
LatvianlvX
LingalalnX
LithuanianltX
LuxembourgishlbX
MacedonianmkX
MalaymsXXX
MalayalammlX
Mandarin ChinesezhXXX
MarathimrX
NepalineX
NorwegiannoXX
PashtopsX
PersianfaX
PolishplXXX
PortugueseptXXX
PunjabipaX
RomanianroXXX
RussianruXXX
SerbiansrX
SindhisdX
SlovakskXXX
SlovenianslX
SomalisoX
SpanishesXXX
SwahiliswX
SwedishsvXXX
TamiltaXXX
TeluguteX
ThaithX
TurkishtrXXX
UkrainianukXXX
UrduurX
VietnameseviXX
WelshcyX

Set voice.language (an ISO 639-1 code) to enforce a specific language. Enforcement is supported only on Eleven Flash v2.5. For the other models you do not need to set a language code. They detect the language from the input text, and providing a code returns an error.