Voices: catalog, languages, and how to pick one
The 26 voices on Call2Me — 16 ElevenLabs Flash v2.5 voices for standard mode and 10 GPT Realtime voices — with languages, gender and how to audition them before going live.
Updated October 6, 2026
A voice is what your agent sounds like. Picking the right one is half
brand decision, half mode choice. This is the full catalog as served by
GET /v1/voices.
An earlier version of this page listed OpenAI and Google Wavenet voices, "40+" voices and self-serve cloning. None of that is available; the page was rewritten from the live catalog on 6 October 2026.
Two voice modes
| Mode | Voices | Engine | Cost |
|---|---|---|---|
| Standard (default) | 16 | Deepgram STT → your LLM → ElevenLabs Flash v2.5 | Included in the per-minute rate |
| Realtime | 10 | GPT speech-to-speech | +$0.25/min |
If ElevenLabs is unreachable, standard-mode calls fall back to a Cartesia voice automatically so the call is never silent. You cannot pick Cartesia directly.
Standard voices (ElevenLabs Flash v2.5)
| Voice | Language | Gender | Character |
|---|---|---|---|
| Selin | Turkish | female | Warm, friendly — the default |
| Nergis | Turkish | female | Elegant, conversational |
| Will | Turkish | male | Relaxed, upbeat |
| George | multilingual | male | Warm storyteller |
| Brian | multilingual | male | Deep, calming |
| River | multilingual | neutral | Calm, informative |
| Bill | multilingual | male | Mature, measured |
| Sarah | multilingual | female | Reassuring, calm |
| Alice | multilingual | female | Clear educator |
| Lily | multilingual | female | Soft, confident |
| Eric | multilingual | male | Smooth professional |
| Adam | multilingual | male | Deep, decisive |
| Matilda | multilingual | female | Knowledgeable, positive |
| Jessica | multilingual | female | Bright, warm |
| Daniel | multilingual | male | Formal announcer |
| Chris | multilingual | male | Casual, natural |
Realtime voices (GPT Realtime)
Marin, Cedar, Alloy, Echo, Shimmer, Ash, Ballad, Coral, Sage, Verse — all multilingual. They are optimized for English; test Turkish before going live.
Languages
The agent language picker offers Turkish, American English, British English, German, French, Spanish and Arabic, plus Multilingual. The 13 multilingual voices speak all of them; for Turkish, the three Turkish-native voices sound the most natural.
Auditioning
Dashboard → Agents → Voice: every voice has a play button.
Programmatically:
# Catalog, filterable by language, provider and gender
curl "https://api.call2me.app/v1/voices?language=tr-TR" \
-H "Authorization: Bearer YOUR_API_KEY"
# MP3 sample of one voice
curl -X POST https://api.call2me.app/v1/voices/elevenlabs-CFzl8WvPQ92EKMBs2hpS/preview \
-H "Authorization: Bearer YOUR_API_KEY" \
-H "Content-Type: application/json" \
-d '{"text": "Merhaba, size nasıl yardımcı olabilirim?"}' --output selin.mp3
Multilingual voices match any language filter.
Picking a voice
- Language. One language → a native voice if there is one (Turkish), otherwise any multilingual voice. Several languages on one agent → Multilingual agent language + a multilingual voice.
- Brand vs. utility. Warm and expressive (George, Jessica) for hospitality and sales; neutral and clear (Daniel, River) for transactional support.
- Mode. Standard covers most use cases. Realtime when the conversation must feel fully fluid and the extra $0.25/min is worth it.
What's next
Frequently asked
Q.How many voices are available?
26: 16 ElevenLabs Flash v2.5 voices (3 Turkish-native, 13 multilingual) for standard mode, and 10 GPT Realtime voices for realtime mode. GET /v1/voices returns the live catalog.
Q.Can I clone my own voice?
Not self-serve. Voices are synthesized through the platform's ElevenLabs account, so a voice cloned in your own ElevenLabs account is not reachable, and an unknown voice_id falls back to a default voice. If you need a custom voice, contact us.
Q.Which voice is fastest?
All standard-mode voices use ElevenLabs Flash v2.5, the low-latency model, so speed does not differ between them. Realtime voices skip separate STT/TTS entirely and feel the most fluid, at +$0.25/min.
Q.Can a single agent speak multiple languages?
Yes. Set the agent language to Multilingual and pick one of the 13 multilingual ElevenLabs voices; the same voice character follows the caller's language.