Home
Jobs
Events
Blogs
Library
contact us
[
Stack Layer
]
Voice models that turn the agent’s reply back into natural speech.
All layers
Submit a company
UNIO EDITORS’ PICKS
#1
Cartesia
Ultra-low-latency text-to-speech models built for real-time voice agents.
#2
ElevenLabs
Expressive multilingual text-to-speech and voice cloning, plus a conversational agents platform.
#3
Smallest.ai
Ultra-low-latency TTS (sub-100ms) and voice cloning.
Orchestration & Agent Platforms
Dubbing and audio localization.
India-ready
Speech-to-Text
IIT Bombay-led govt consortium; speech model Shrutam.
+LLM +TTS
Text-to-Speech
Fast, accurate speech-to-text APIs built for real-time voice agents.
+TTS
+STT +Orchestration
Open-source Maya1 TTS model; $1.9M raised.
Sovereign Indic LLMs plus STT/TTS models.
+TTS +LLM
Voice AI cloud: one API to run any speech-to-text, text-to-speech or voice-cloning model across 20+ regions, with data residency built in.
In-house Vani TTS behind unlimited-calling plans.
+Orchestration