Voices
Available voices for text-to-speech synthesis.
Indic Voices
bulbul:v3 provides 37 voices spanning 11 Indian languages (bn-IN, en-IN, gu-IN, hi-IN, kn-IN, ml-IN, mr-IN, od-IN, pa-IN, ta-IN, te-IN). Pass the voice ID as the voice parameter and the target language. The default voice is shubh. An unrecognized voice falls back to the recommended voice for the request's language (for example ratan for en-IN, ta-IN, mr-IN and gu-IN, shubh for the rest).
shubh · aditya · ritu · priya · neha · rahul · pooja · rohan · simran · kavya
amit · dev · ishita · shreya · ratan · varun · manan · sumit · roopa · kabir
aayan · ashutosh · advait · anand · tanya · tarun · sunny · mani · gokul · vijay
shruti · suhani · mohit · kavitha · rehan · soham · rupaliPreview every voice in the Playground.
bulbul:v4-flash provides 222 persona voices across 9 Indian languages (hi-IN, en-IN, bn-IN, gu-IN, kn-IN, mr-IN, pa-IN, ta-IN, te-IN) plus Hinglish (English-Hindi code-mixed) personas. Voice IDs follow <name>_<language>_<style>, for example simran_en_customer or aparna_hi_customer. The default voice is shubh_enhi_ads. The bulbul:v3 names above (shubh, priya and the rest) are not valid on this model, and ml-IN and od-IN are not available on it; use bulbul:v3 for those.
Default voice per language:
| Language | Default voice |
|---|---|
hi-IN, en-IN | shubh_enhi_ads |
bn-IN | arnab_bn_conversation |
gu-IN | bhavik_gu_conversation |
kn-IN | chaitra_kn_conversation |
mr-IN | ishita_mr_conversational |
pa-IN | anand_pa_conversation |
ta-IN | gokul_ta_narration |
te-IN | kavitha_te_conversation |
The full list, grouped by language. GET /api/v1/models also returns it as voice_names on the bulbul:v4-flash entry.
Hindi (hi-IN) — 116 voices
aayan_hi_conversational · amit_hi_conversational · ashutosh_hi_conversational · kabir_hi_conversational
kavya_hi_conversational · manan_hi_conversational · rahul_hi_conversational · sumit_hi_conversational
aditya_hi_conversational · aditya_hi_sales · anand_hi_documentary · anand_hi_news
aparna_hi_customer · aparna_hi_kyc · ashok_hi_character · ashok_hi_news
chhavi_hi_kids · ishita_hi_ads · ishita_hi_edtech · ishita_hi_banking
ishita_hi_ads_informal · ishita_hi_devotional · ishita_hi_numbers · ishita_hi_social
kunal_hi_kids · mahesh_hi_documentary · mani_hi_devotional · mani_hi_conversational
mohit_hi_conversational · nachiket_hi_devotional · priya_hi_recovery · ratan_hi_latenight
ratan_hi_customer_expressive · ratan_hi_documentary · ratan_hi_devotional · ratan_hi_recovery
ratan_hi_social · ratan_hi_sports · ratan_hi_latenight_warm · rehan_hi_social
ritu_hi_customer_utility · ritu_hi_kids · ritu_hi_conversation · ritu_hi_customer
ritu_hi_edtech · ritu_hi_ads_formal · ritu_hi_banking · ritu_hi_ads_informal
ritu_hi_insurance · ritu_hi_edtech_bot · ritu_hi_medical · ritu_hi_sales
ritu_hi_reels · ritu_hi_social · ritu_hi_customer_warm · ritu_hi_social_lively
roopa_hi_companion · roopa_hi_narration · roopa_hi_recovery · roopa_hi_market
roopa_hi_conversational · sanchita_hi_assistant · sanchita_hi_edtech · sanchita_hi_banking
sanchita_hi_feedback · sanchita_hi_ads_formal · sanchita_hi_ads_informal · sanchita_hi_interview
sanchita_hi_romantic · sanchita_hi_market · sanchita_hi_social · sanchita_hi_kyc
sarika_hi_conversation · shalini_hi_companion · shalini_hi_social · shreya_hi_conversational
shreya_hi_news · shruti_hi_edtech · shubh_hi_customer · shubh_hi_ecomm
shubh_hi_stories_mixed · shubh_hi_devotional · shubh_hi_ads · shubh_hi_recovery
shubh_hi_stories_dramatic · simran_hi_assistant · simran_hi_narration · simran_hi_automobile
simran_hi_conversation · simran_hi_news_breaking · simran_hi_social_energetic · simran_hi_social_excited
simran_hi_latenight · simran_hi_news · simran_hi_recovery · simran_hi_sales
suchitra_hi_ecomm · suhani_hi_social · sunny_hi_ads · sunny_hi_reels
tarun_hi_conversational · tarun_hi_sales · chaitra_hi_customer · shilpa_hi_narration
tanya_hi_narration · aarti_hi_customer · advait_hi_character · aryaman_hi_ads
chirag_hi_social · girish_hi_devotional · mukul_hi_ads · mukul_hi_suspense
suman_hi_companion · vaibhav_hi_social · vandana_hi_ecomm · vipul_hi_socialEnglish (en-IN) — 62 voices
aditi_en_stories · aparna_en_companion · aparna_en_edtech · ashwin_en_sports
ashwin_en_sports_energetic · chandrika_en_stories · dev_en_recovery · dev_en_conversational
deven_en_conversation · ishita_en_customer · ishita_en_medical · ishita_en_numbers
ishita_en_social · ishita_en_stories · kalpit_en_edtech · nachiket_en_ads
neha_en_customer · neha_en_latenight · nupur_en_kids · ojas_en_social
ritu_en_edtech · ritu_en_latenight · ritu_en_medical · ritu_en_reels
rohan_en_recovery · roopa_en_conversational · rustom_en_suspense · sanchita_en_companion
sanchita_en_insurance · sanchita_en_recovery · sanchita_en_market · sanchita_en_social
shabana_en_edtech · shalini_en_companion · shalini_en_customer · shubh_en_narration
shubh_en_numbers · shubh_en_ads · shubh_en_recovery · shubh_en_audiobook
shubh_en_narration_gentle · shubh_en_sports · simran_en_narration · simran_en_automobile
simran_en_conversation · simran_en_customer · simran_en_edtech · simran_en_edtech_bot
simran_en_sales · simran_en_recovery · simran_en_ads · simran_en_therapist
sunny_en_social · varun_en_ads · varun_en_suspense · zarina_en_conversation
amelia_en_conversational · sophia_en_conversational · girish_en_documentary · girish_en_devotional
payal_en_edtech · sarang_en_narrationHinglish, English-Hindi code-mixed (hi-IN / en-IN) — 13 voices
ishita_enhi_companion · ishita_enhi_customer · ishita_enhi_customer_expressive · sanchita_enhi_companion
shalini_enhi_companion · shalini_enhi_customer · shubh_enhi_companion · shubh_enhi_ads
shubh_enhi_banking · simran_enhi_companion · simran_enhi_customer · simran_enhi_banking_expressive
sunny_enhi_customerMarathi (mr-IN) — 9 voices
ishita_mr_conversational · mrunal_mr_narration · neha_mr_narration · nilesh_mr_conversation
ritu_mr_insurance · ritu_mr_narration · rupali_mr_stories · soham_mr_narration
mukul_mr_storiesKannada (kn-IN) — 4 voices
chaitra_kn_conversation · chaitra_kn_narration · chetan_kn_conversation · suchitra_kn_narrationPunjabi (pa-IN) — 4 voices
anand_pa_conversation · anand_pa_customer · harpreet_pa_narration · jaspal_pa_bankingTamil (ta-IN) — 4 voices
gokul_ta_narration · aravind_ta_ads · aravind_ta_suspense · vijay_ta_narrationTelugu (te-IN) — 4 voices
kavitha_te_conversation · kavitha_te_narration · pooja_te_conversation · tarun_te_narrationGujarati (gu-IN) — 3 voices
bhavik_gu_conversation · pooja_gu_conversational · pooja_gu_customerBengali (bn-IN) — 3 voices
arnab_bn_conversation · roopa_bn_conversational · bimal_bn_suspensegnani-timbre-v3.0 (Gnani Timbre v3) provides 12 voices. Every voice speaks all 10 Indian languages (hi-IN, en-IN, ta-IN, te-IN, kn-IN, ml-IN, mr-IN, pa-IN, bn-IN, gu-IN) plus auto and Hinglish hi-en. The default voice is Shivani; an unrecognized voice falls back to the default. The text also accepts expressive audio tags.
| Voice | Style | Character |
|---|---|---|
Shivani (default) | Feminine | Warm, patient, empathy-first voice for inbound customer support |
Lakshmi | Feminine | Expressive, warm voice for storytelling and e-learning narration |
Kalpana | Feminine | Empathetic, composed voice for customer care and insurance |
Chetana | Feminine | Gentle, empathetic voice for healthcare and telecom customer service |
Vasudha | Feminine | Expressive, composed voice for narration and e-learning |
Riyaa | Feminine | Confident, lively voice for outbound sales and insurance |
Tarika | Feminine | Warm, reassuring voice for banking and insurance customer care |
Avinash | Masculine | Professional, firm voice with deliberate pacing for collections and BFSI |
Shubham | Masculine | Energetic, confident voice for outbound sales and lead qualification |
Prakash | Masculine | Steady, deliberate voice for payment recovery and BFSI |
Devansh | Masculine | Energetic, young conversational voice for sales and fintech |
Yashraj | Masculine | Warm narrative voice for e-learning and IVR narration |
Cartesia Voices
sonic-3.6 (Cartesia Sonic 3.6) speaks 44 languages with native-quality Hindi and Hinglish. Pass a featured handle (skylar, default) or any public Cartesia voice UUID as voice, plus a base ISO language code (en, hi). An unrecognized non-UUID handle falls back to skylar.
The live public library is paginated — do not treat the 16 featured aliases as the full set:
curl "https://api.callmissed.com/api/v1/models/sonic-3.6/voices?q=hindi&limit=50"
curl "https://api.callmissed.com/v1/audio/voices?model=sonic-3.6&q=skylar" \
-H "Authorization: Bearer cm_YOUR_KEY"GET /api/v1/models/sonic-3.6/voices/{id}/preview streams Cartesia's own pre-recorded sample (no synthesis, no credit charge). Neither listing endpoint synthesizes or bills.
Query params (both listing endpoints): q (search, max 200 chars), language (base code such as hi), gender (masculine / feminine / gender_neutral; male / female also accepted), limit (1–100, default 50), starting_after (a voice id from the previous page). GET /v1/audio/voices takes model=sonic-3.6 (the default) or any eleven_* model (ElevenLabs library, paged with page); any other model returns 400.
{
"object": "list",
"model": "sonic-3.6",
"data": [
{
"id": "voice-uuid",
"handle": "skylar",
"name": "Skylar",
"gender": "feminine",
"language": "en",
"locale": "en-US",
"accent": "",
"tagline": "",
"description": "",
"has_preview": true,
"preview": "/api/v1/models/sonic-3.6/voices/voice-uuid/preview"
}
],
"has_more": true,
"next_page": "…"
}handle is set only for the featured aliases below (otherwise null); pass either the id or the handle as voice on POST /v1/audio/speech. preview is null when a voice has no sample. When has_more is true, pass the last voice's id as starting_after to fetch the next page.
If the upstream library is briefly unreachable, the first page answers 200 with the featured aliases below and an extra "degraded": true field instead of failing, so a voice picker still renders and every returned voice remains usable for synthesis. The field is absent on a normal response — treat its presence as "this is the short list, retry later for the full library". A request carrying starting_after is not degraded: pagination returns 503 so you keep the page you already have.
Featured aliases (stable handles, also valid UUIDs in the library):
skylar · daniel · jacqueline · katie · cathy · caroline · ronald · carson · jameson
gemma · archie · riya · arushi · siya · parvati · kabirriya, arushi, siya, parvati, and kabir are native Hindi speakers (pair with "language": "hi"). Browse the full library in the console Voices page, the Playground, or Talk.
Other TTS providers also expose voices via the same
POST /v1/audio/speechendpoint — deepgram-aura-2 (90 voices across English, Spanish, German, French, Dutch, Italian, and Japanese, defaultthalia), deepgram-aura-1 (12 legacy English voices at half the Aura-2 rate, defaultasteria), and gpt-4o-mini-tts (alloy,ash,ballad,coral,echo,fable,nova,onyx,sage,shimmer,verse,marin,cedar). See Credits & Rate Limits for per-model pricing.
ElevenLabs Voices
The nine ElevenLabs models — eleven_v4, eleven_v4_turbo, eleven_v3, eleven_v3_conversational, eleven_multilingual_v2, eleven_flash_v2_5, eleven_turbo_v2_5, eleven_flash_v2 and eleven_turbo_v2 (see Text to Speech) — share one voice library. Pass a premade name from the table below (case-insensitive; sarah is the default) or the 20-character id of any voice in the voice library (search it with the endpoint below) as voice. An unrecognised name falls back to sarah. Every premade voice speaks every language its model supports; the accent is the voice's own.
| Voice | Gender | Accent | Character |
|---|---|---|---|
sarah (default) | Feminine | American | Mature, reassuring, confident |
laura | Feminine | American | Enthusiast, quirky attitude |
jessica | Feminine | American | Playful, bright, warm |
matilda | Feminine | American | Knowledgeable, professional |
bella | Feminine | American | Professional, bright, warm |
alice | Feminine | British | Clear, engaging educator |
lily | Feminine | British | Velvety actress |
river | Neutral | American | Relaxed, neutral, informative |
roger | Masculine | American | Laid-back, casual, resonant |
charlie | Masculine | Australian | Deep, confident, energetic |
george | Masculine | British | Warm, captivating storyteller |
callum | Masculine | American | Husky trickster |
harry | Masculine | American | Fierce warrior |
liam | Masculine | American | Energetic, social media creator |
will | Masculine | American | Relaxed optimist |
eric | Masculine | American | Smooth, trustworthy |
chris | Masculine | American | Charming, down-to-earth |
brian | Masculine | American | Deep, resonant and comforting |
daniel | Masculine | British | Steady broadcaster |
adam | Masculine | American | Dominant, firm |
bill | Masculine | American | Wise, mature, balanced |
GET /api/v1/models?category=tts returns these as voice_details for each eleven_* model.
Searching the ElevenLabs voice library
Beyond the 21 premade voices, more than 17,000 library voices are searchable, about 7,000 of them Hindi and 1,000 Indian English. Results are paginated and include only voices billed at the model's standard per-character rate. Listing does not synthesize or bill.
curl "https://api.callmissed.com/api/v1/models/eleven_flash_v2_5/voices?language=hi&gender=female&limit=50"
curl "https://api.callmissed.com/v1/audio/voices?model=eleven_flash_v2_5&q=customer%20care" \
-H "Authorization: Bearer cm_YOUR_KEY"Query params: q (search names, descriptions and voice IDs, max 200 chars), language (base code such as hi), gender (feminine / masculine / gender_neutral; female / male / neutral also accepted), limit (1–100, default 50), and page (0-based). Any eleven_* model id works in the path; every model shares the same library.
{
"object": "list",
"model": "eleven_flash_v2_5",
"data": [
{
"id": "0PvBRbbZjTJPB3AadF6T",
"handle": null,
"name": "Anika - Quick Pace Customer Care",
"gender": "feminine",
"language": "hi",
"locale": "hi-IN",
"accent": "",
"age": "young",
"tagline": "pleasant, conversational",
"description": "…",
"has_preview": true,
"preview": "https://…/preview.mp3"
}
],
"has_more": true,
"next_page": 1,
"total": 7178
}Pass id as voice on POST /v1/audio/speech or as an agent's voice. preview is a direct link to the voice's own recorded sample. A premade voice comes back with its handle (for example sarah).
- Paging: the first page (
page=0) without alanguagefilter starts with the matching premade voices. Whenhas_moreistrue, requestnext_pageto get the next page. - When the library is briefly unreachable: the first page still answers
200, with only the premade voices and"degraded": true. Later pages return503.
MAI Voice (preview)
mai-voice-2.1 ($0.2292 / 10K chars) and mai-voice-2.1-flash ($0.1563 / 10K chars) are Microsoft AI's text-to-speech models. Both are preview. They share the same 97 voices across 23 languages: cs, da, de, en, es, fi, fr, hi, hu, id, it, ko, nb, nl, pl, pt, ro, ru, sv, th, tr, vi, zh. They work on POST /v1/audio/speech and in voice agents and voice sessions (tts_model). mai-voice-2.1-flash is built for low latency and is the pick for live calls; mai-voice-2.1 is the most expressive and suits narration and long-form audio.
Voice handles are lowercase <locale>-<name> (for example hi-in-kavya) and case-insensitive. A voice speaks only its own locale. With no voice, the first voice for the request language is used (hi gives hi-in-arjun), otherwise en-in-priya. Input is plain text: SSML and tags are not supported, and the text is read as written. The one voice-agent control is tts_speed, 0.5 to 2.0. GET /api/v1/models?category=tts returns every voice in voice_names and voice_details.
| Voice | Gender | Locale |
|---|---|---|
hi-in-arjun (Hindi default) | Masculine | Hindi |
hi-in-dhruv | Masculine | Hindi |
hi-in-grant | Masculine | Hindi |
hi-in-harper | Feminine | Hindi |
hi-in-kavya | Feminine | Hindi |
hi-in-priya | Feminine | Hindi |
en-in-priya (default) | Feminine | English (India) |
en-in-dhruv | Masculine | English (India) |
They are not processed in India, so an agent with hosting: "india" cannot use them. They are paid-plan only, are never a default, cannot be used in tts_fallbacks, and are not offered on the public voice demo.
Preview models may change or be retired at short notice. If one is retired, its ID keeps working and is answered by a named successor, as described in Retired and deprecated models, so an agent that uses it keeps working but may sound different. Use a generally available model for production-critical agents.
Voice session voices
Six more text-to-speech models speak only inside a voice agent: set one as tts_model on an agent or a voice session. POST /v1/audio/speech returns 400 for them. The section also covers deepgram-flux-tts, which is valid on POST /v1/audio/speech as well. Pass the voice as voice (a handle from the tables below, case-insensitive) and a base ISO code as language.
Prices are per 10K characters. GET /api/v1/models?category=tts returns voice_details (gender, accent and description per voice), language_codes, supports_ssml, supports_delivery_tags and audio_tags for each of them, so a picker can be built from the API.
Markup support differs per model. A voice agent's prompt teaches its model the right form automatically; when you send text yourself, use this:
| Model | Delivery directions | Sounds | Pauses | Everything else |
|---|---|---|---|---|
inworld-tts-2 | <expression value="gently warm"/> before a sentence, written in English | <sound value="laugh"/>: laugh, sigh, breathe, clear throat, cough | <break time="500ms"/> | — |
inworld-tts-2-flash | Not performed (ignored) | Same as TTS-2 | <break time="500ms"/> | — |
fish-audio-drama-3 | <expression value="warm"/> | <sound value="chuckling"/>: laughing, chuckling, sighing | — | Write amounts and dates in words: the ₹ sign is read as dollars |
rime-coda | — | — | — | Plain text; tags and SSML are read aloud. spell(AB79X) spells a code |
deepgram-flux-tts | — | — | — | Plain text; tags and SSML are read aloud, and a {pause} control stops the audio |
gradium-tts-beta | — | — | — | Plain text; tags are read aloud and ₹ is read as euros |
microsoft-neural-tts | — | — | — | Plain text only: no SSML, no audio tags, no delivery tags. It reads rupee amounts and dates itself |
On fish-audio-drama-3, gradium-tts-beta, rime-coda and deepgram-flux-tts, voice agents rewrite amounts, slash dates (day first) and long digit runs into words before speaking, because these models misread them otherwise. Set tts_normalize to false on the agent to turn this off.
deepgram-flux-tts — Deepgram Flux TTS
Deepgram's conversational TTS built for voice agents, $0.4688 / 10K chars. 266 voices in 42 languages, and the only Deepgram model here that is also valid on POST /v1/audio/speech (see Text to Speech). 149 voices in 10 languages (including Hindi) are generally available; 117 voices in 32 languages (including Tamil, Telugu, Kannada, Marathi, Gujarati, Bengali and Urdu) are Early Access. Pass the voice as a bare name (ajay); each voice belongs to one language, so the voice sets the language. Leave voice out to get the default for language (priya for English, or a language Flux has no voice for).
Early Access languages read numbers and dates as written, so spell them out ("two thousand twenty-six", not "2026"); the speed control works for English voices only. Every voice in GET /v1/models/deepgram-flux-tts carries a language field.
| Code | Language | Voices | Status | Default voice |
|---|---|---|---|---|
en | English | 36 | GA | priya |
nl | Dutch | 13 | GA | anouk |
fr | French | 12 | GA | arnaud |
de | German | 13 | GA | benjamin |
hi | Hindi | 14 | GA | ajay |
it | Italian | 11 | GA | angela |
ja | Japanese | 4 | GA | ayumi |
pt | Portuguese (Brazil) | 18 | GA | bianca |
ru | Russian | 9 | GA | andrei |
es | Spanish | 19 | GA | aina |
bn | Bengali | 6 | Early Access | afsana |
bg | Bulgarian | 2 | Early Access | desislava |
ca | Catalan | 2 | Early Access | jordi |
zh | Chinese (Mandarin) | 14 | Early Access | bo |
hr | Croatian | 7 | Early Access | dalibor |
cs | Czech | 1 | Early Access | jakub |
da | Danish | 3 | Early Access | pernille |
et | Estonian | 1 | Early Access | triin |
fi | Finnish | 2 | Early Access | kirsi |
el | Greek | 1 | Early Access | kostas |
gu | Gujarati | 3 | Early Access | bhavna |
hu | Hungarian | 4 | Early Access | bence |
id | Indonesian | 5 | Early Access | dewi |
kn | Kannada | 1 | Early Access | manjunath |
ko | Korean | 6 | Early Access | eunji |
lv | Latvian | 2 | Early Access | agnese |
lt | Lithuanian | 2 | Early Access | goda |
ms | Malay | 2 | Early Access | haziq |
mr | Marathi | 3 | Early Access | pooja |
no | Norwegian | 6 | Early Access | erlend |
pl | Polish | 6 | Early Access | bartosz |
ro | Romanian | 2 | Early Access | cosmin |
sk | Slovak | 2 | Early Access | dominika |
sl | Slovenian | 2 | Early Access | gregor |
sv | Swedish | 7 | Early Access | axel |
tl | Tagalog | 5 | Early Access | jerome |
ta | Tamil | 4 | Early Access | gokul |
te | Telugu | 3 | Early Access | pavan |
tr | Turkish | 7 | Early Access | dilek |
uk | Ukrainian | 2 | Early Access | denys |
ur | Urdu | 2 | Early Access | ambreen |
vi | Vietnamese | 2 | Early Access | hoa |
The English voices with a full description:
| Voice | Gender | Accent | Description |
|---|---|---|---|
priya (default) | Feminine | Indian | Self-assured and reassuring, with an empathetic, calm professionalism |
meena | Feminine | Indian | Patient and reassuring, with a calm, professional empathy |
naveen | Masculine | Indian | Knowledgeable and composed, with a clear, caring professionalism |
hannah | Feminine | American | Articulate and self-assured, with a thoughtful, pleasant manner |
alexis | Feminine | American | Poised and articulate, with a calm, compassionate tone |
cole | Masculine | American | Friendly and lively, with an engaging, energetic style |
miles | Masculine | American | Calm and sincere, with a clear, confident professional voice |
sienna | Feminine | American | Polished and steady, with a warm, caring delivery |
cliff | Masculine | American | Low-pitched and self-assured, with a calm, slightly gravelly quality |
brooke | Feminine | American | Quick-paced and upbeat, with an intelligent, self-possessed feel |
haley | Feminine | American | Professional and calm, with an empathetic, caring tone |
kit | Masculine | British | Lively but composed, with a helpful, thoughtful streak |
colin | Masculine | British | Warm and dependable, with an authoritative yet approachable presence |
gemma | Feminine | British | Cheerful and kind, with an approachable, caring warmth |
sharon | Feminine | Australian | Formal yet relaxed, with a calm, self-assured manner |
kai | Masculine | Singaporean | Knowledgeable and calm, with a caring, precise professional style |
All 266 voice names are listed under Flux TTS voices by language.
rime-coda — Rime Coda
Rime's flagship conversational TTS, $0.5208 / 10K chars. 8 languages: en, hi, es, fr, de, ja, pt, ar. 13 voices; nadi and taru are native Hindi speakers. The default is nadi.
| Voice | Gender | Accent | Description |
|---|---|---|---|
nadi (default) | Feminine | Indian | A balanced female voice. Native Hindi |
taru | Masculine | Indian | A balanced, professional sounding native Hindi speaker |
astra | Feminine | American | Bright, lively California voice, quick and expressive |
luna | Feminine | American | A bright, lively voice, energetic with a casual core |
lyra | Feminine | American | A warm young American voice, bright and lively |
celeste | Feminine | American | Chill Gen-Z American voice |
clementine | Feminine | American | A warm, lively twenty-something American voice, polished and energetic |
eyre | Feminine | American | A warm, friendly American voice, calm and easy to listen to |
vespera | Feminine | American | A friendly, lively young Californian voice, bright and adaptable |
bancroft | Masculine | American | A composed American voice, strong on professional and formal contexts |
cupola | Masculine | American | A confident, energetic American voice, professional with warmth |
lawton | Masculine | American | A warm American voice, calm and friendly throughout |
marlu | Masculine | Australian | An older Australian voice; low, mature, unhurried |
inworld-tts-2 and inworld-tts-2-flash — Inworld TTS-2
Inworld's newest TTS. Every voice speaks every language: 94 base codes, including Hindi (hi), Marathi (mr), Bengali (bn), Tamil (ta), Telugu (te), Gujarati (gu), Kannada (kn), Malayalam (ml), Punjabi (pa), Urdu (ur) and Odia (or). inworld-tts-2 is $0.2604 / 10K chars; inworld-tts-2-flash is the faster, lower-cost model at $0.1563 / 10K chars. Both take the same 14 voices; the default is riya. Inworld publishes gender and a description for only some voices, so the others show a dash.
| Voice | Gender | Accent | Description |
|---|---|---|---|
riya (default) | — | Indian | Hindi voice |
manoj | — | Indian | Hindi voice |
priya | — | — | — |
saanvi | — | — | — |
aanya | — | — | — |
shreya | — | — | — |
aarav | — | — | — |
arjun | — | — | — |
ashley | Feminine | American | Warm, natural American voice |
olivia | Feminine | British | Upbeat, friendly British voice |
edward | Masculine | American | Fast-talking, emphatic American voice |
diego | Masculine | Mexican | Soothing, gentle Mexican voice |
sarah | — | American | — |
lauren | — | American | — |
fish-audio-drama-3 — Fish Audio Drama 3
Fish Audio's expressive TTS, $0.5208 / 10K chars. 13 languages (en, zh, ja, de, fr, es, ko, ar, ru, nl, it, pl, pt), no Indian languages. 7 voices, default selene.
| Voice | Gender | Accent | Description |
|---|---|---|---|
selene (default) | Feminine | American | Meditative voice |
laura | Feminine | American | Confident narrator |
adrian | Masculine | American | Friendly, casual American voice |
hannah | — | American | Conversational specialist |
ethan | — | American | Curious explainer |
sarah | — | American | Engaged speaker |
jordan | — | American | Motivational speaker |
gradium-tts-beta — Gradium TTS
Gradium's newest TTS, $0.50 / 10K chars. 5 languages (en, fr, de, es, pt), no Indian languages. 13 voices, default harper.
| Voice | Gender | Accent | Description |
|---|---|---|---|
harper (default) | Feminine | American | Modern, confident and friendly voice with a standard American accent |
zoey | Feminine | American | Playful, upbeat Gen Z energy with a standard American accent |
sunnie | Feminine | American | A joyful young American voice, warm and bubbly, keeping support calls bright and light |
brooklyn | Feminine | American | A warm, effusive young American voice that makes customers feel at home |
sterling | Masculine | American | A warm, energetic American adult voice with theatrical flair |
garrett | Masculine | American | A smooth, low-pitched American adult voice with easy confidence |
marcus | Masculine | American | A high-energy, resonant American adult voice with unshakeable conviction |
tilly | Feminine | British | A bright, welcoming British voice with the warmth of a great receptionist |
maeve | Feminine | British | A sparky, attentive British voice that gets to the point with a smile |
archie | Masculine | British | A bright, welcoming British voice, made for front-line support |
toby | Masculine | British | A sparky, attentive British voice for fast, efficient support |
aoife | Feminine | Irish | Bright and engaging Irish English voice with lively, natural pacing |
cormac | Masculine | Irish | Bright and warm Irish English voice with cheerful, natural pacing |
microsoft-neural-tts — Microsoft Neural voices
Microsoft's neural text-to-speech, processed in India, $0.1563 / 10K chars. 57 voices across 13 languages: as, bn, en, gu, hi, kn, ml, mr, or, pa, ta, te, ur. The default is en-in-neerja. Input is plain text, with no SSML, audio tags or delivery tags, and the model normalises rupee amounts and dates itself. The one control is speaking speed, tts_speed from 0.5 to 2.0 (default 1.0).
A voice speaks only its own locale (the handle starts with it, for example hi-in-swara speaks Hindi), except the four Indic voices (en-in-aartiindic, en-in-arjunindic, en-in-neerjaindic, en-in-prabhatindic), which speak English (India) plus Hindi, Tamil, Telugu, Gujarati, Marathi, Assamese, Punjabi, Odia, Malayalam, Kannada and Bengali. Gender is the value Microsoft publishes for each voice; Microsoft publishes a description for the English (US) and English (UK) voices only.
| Voice | Gender | Locale | Description |
|---|---|---|---|
en-in-neerja (default) | Feminine | English (India) | — |
en-in-aartiindic | Feminine | English (India) | Speaks English (India) plus 11 Indian languages |
en-in-arjunindic | Masculine | English (India) | Speaks English (India) plus 11 Indian languages |
en-in-neerjaindic | Feminine | English (India) | Speaks English (India) plus 11 Indian languages |
en-in-prabhatindic | Masculine | English (India) | Speaks English (India) plus 11 Indian languages |
en-in-aarav | Masculine | English (India) | — |
en-in-aashi | Feminine | English (India) | — |
en-in-aarti | Feminine | English (India) | — |
en-in-arjun | Masculine | English (India) | — |
en-in-ananya | Feminine | English (India) | — |
en-in-kavya | Feminine | English (India) | — |
en-in-kunal | Masculine | English (India) | — |
en-in-prabhat | Masculine | English (India) | — |
en-in-rehaan | Masculine | English (India) | — |
hi-in-swara | Feminine | Hindi | — |
hi-in-madhur | Masculine | Hindi | — |
hi-in-aarav | Masculine | Hindi | — |
hi-in-ananya | Feminine | Hindi | — |
hi-in-aarti | Feminine | Hindi | — |
hi-in-arjun | Masculine | Hindi | — |
hi-in-kavya | Feminine | Hindi | — |
hi-in-kunal | Masculine | Hindi | — |
hi-in-rehaan | Masculine | Hindi | — |
as-in-yashica | Feminine | Assamese | — |
as-in-priyom | Masculine | Assamese | — |
bn-in-tanishaa | Feminine | Bengali | — |
bn-in-bashkar | Masculine | Bengali | — |
gu-in-dhwani | Feminine | Gujarati | — |
gu-in-niranjan | Masculine | Gujarati | — |
kn-in-sapna | Feminine | Kannada | — |
kn-in-gagan | Masculine | Kannada | — |
ml-in-sobhana | Feminine | Malayalam | — |
ml-in-midhun | Masculine | Malayalam | — |
mr-in-aarohi | Feminine | Marathi | — |
mr-in-manohar | Masculine | Marathi | — |
or-in-subhasini | Feminine | Odia | — |
or-in-sukant | Masculine | Odia | — |
pa-in-vaani | Feminine | Punjabi | — |
pa-in-ojas | Masculine | Punjabi | — |
ta-in-pallavi | Feminine | Tamil | — |
ta-in-valluvar | Masculine | Tamil | — |
te-in-shruti | Feminine | Telugu | — |
te-in-mohan | Masculine | Telugu | — |
ur-in-gul | Feminine | Urdu | — |
ur-in-salman | Masculine | Urdu | — |
en-us-ava | Feminine | English (US) | Pleasant, Caring, Friendly. Suited to chat, documentary |
en-us-andrew | Masculine | English (US) | Confident, Authentic, Warm. Suited to chat, advertisement |
en-us-emma | Feminine | English (US) | Cheerful, Light-hearted, Casual. Suited to e-learning, chat |
en-us-brian | Masculine | English (US) | Sincere, Calm, Approachable. Suited to podcast, chat |
en-us-jenny | Feminine | English (US) | Sincere, Pleasant, Approachable. Suited to news, assistant |
en-us-guy | Masculine | English (US) | Light-hearted, Whimsical, Friendly. Suited to news, gaming |
en-us-aria | Feminine | English (US) | Crisp, Bright, Clear. Suited to news, e-learning |
en-us-davis | Masculine | English (US) | Soothing, Calm, Smooth. Suited to audiobook, meditation |
en-gb-sonia | Feminine | English (UK) | Gentle, Soft. Suited to narration, news |
en-gb-ryan | Masculine | English (UK) | Bright, Engaging. Suited to assistant, news |
en-gb-libby | Feminine | English (UK) | Crisp, Bright, Clear. Suited to news, e-learning |
en-gb-maisie | Feminine | English (UK) | Crisp, Cheerful, Bright. Suited to audiobook, narration |
Flux TTS voices by language
The managed Voice Agent's Flux engine (tts_engine: "flux") uses the same 266 voices as deepgram-flux-tts, in all 42 languages; it is billed inside the per-minute voice rate. Synthesis is turn-based and prosody carries across turns. The default is priya, an Indian-accented English voice, and with no voice set a call speaks its own language's default voice. An unrecognized voice falls back to that default.
English (36):
alexis · bree · brittany · brooke · bruce · cliff · cole · colin · conor
donovan · drew · elise · gemma · haley · hannah · heather · jack · kai
kelsey · kit · maeve · marcelo · marcus · meena · meghan · miles · naveen
paige · priya · rufus · sean · sharon · sienna · tanner · wade · wesOther languages (language code, then voices):
nl: anouk · bram · dirk · fleur · irene · lotte · luuk · monique · niels · ruben · sanne · veerle · yvonne
fr: arnaud · benoit · brigitte · camille · chloe · florian · julien · mathieu · pierre · thierry · veronique · yasmine
de: benjamin · christian · janine · leonie · magdalena · moritz · nadine · rolf · sonja · sven · tobias · verena · yannick
hi: ajay · anjali · arjun · divya · jyoti · nikhil · pallavi · prachi · rahul · rajesh · ritu · shreya · sunita · vikram
it: angela · chiara · federico · giorgio · matteo · paola · rocco · sofia · stefano · vincenzo · vito
ja: ayumi · hayato · nozomi · sayaka
pt: artur · benedita · bianca · bruno · catarina · claudia · hugo · leandro · marcia · mariana · marta · mateus · miguel · pedro · renato · ricardo · roberto · rodrigo
ru: andrei · kirill · maxim · mikhail · natalia · nikolai · olga · polina · vladimir
es: aina · alfredo · carmen · cecilia · duvan · eduardo · ernesto · fernanda · fiorella · gustavo · jaime · martin · mauricio · patricia · pilar · rafael · tomas · valentina · ysabel
bn: afsana · ishrat · mamun · sabbir · shuvo · sumaiya
bg: desislava · gergana
ca: jordi · judit
zh: bo · chiaying · haoran · jian · jie · kuanyu · lijuan · manlok · shufen · wenbo · wenjing · yawen · yuting · zhen
hr: dalibor · danijela · domagoj · jasna · josipa · mihaela · vedran
cs: jakub
da: pernille · rasmus · rikke
et: triin
fi: kirsi · teemu
el: kostas
gu: bhavna · hetal · jignesh
hu: bence · eszter · kinga · tibor
id: dewi · fajar · fitri · rizky · teguh
kn: manjunath
ko: eunji · hyunwoo · mikyung · minjun · soojin · sungmin
lv: agnese · edgars
lt: goda · neringa
ms: haziq · izzah
mr: pooja · pratik · sneha
no: erlend · espen · ida · signe · silje · vegard
pl: bartosz · ewa · justyna · kacper · konrad · zuzanna
ro: cosmin · ioana
sk: dominika · juraj
sl: gregor · katja
sv: axel · charlotta · frida · ludvig · oskar · pontus · wilma
tl: jerome · jomar · liezel · noel · trisha
ta: gokul · kavya · sowmya · vinoth
te: pavan · ravi · sravani
tr: dilek · ecem · emre · gizem · mehmet · serkan · zeynep
uk: denys · solomiya
ur: ambreen · usman
vi: hoa · minhThe model exposes no expressive/emotion/style controls and does not interpret SSML.