









Gradium is a voice AI platform for developers that provides ultra-low-latency text-to-speech, speech-to-text, speech-to-speech translation, live translation, voice cloning, and on-device text-to-speech through a unified API. It is designed for building real-time voice agents and conversational applications, with expressive speech generation, accurate transcription, multilingual support, speaker cloning, bidirectional WebSocket streaming, scalable concurrency, and deployment options including cloud, dedicated instances, self-hosted, and on-premises infrastructure.
Create a Gradium account, choose a plan or start with the free tier, and access the Studio or API. Developers can integrate the Python or Rust SDKs, connect through bidirectional WebSocket streaming, and use Gradium with frameworks such as LiveKit, Pipecat, Gradbot, and other voice-agent platforms. Send text for speech generation, audio for transcription, or voice and audio streams for real-time translation and conversational agents.

Free
$0 per month
45,000 included credits, Studio and API access, 2 TTS concurrency, 3 STT concurrency, 2 S2S concurrency, 5 instant voice clones, no Pro Voice Clones, and no commercial use.
XS
$13 per month
225,000 included credits, approximately 5 hours of TTS, 21 hours of STT, 16 hours of STT translation, and 2 to 4 hours of speech-to-speech translation depending on usage. Includes 1,000 instant voice clones and commercial use.
S
$43 per month
900,000 included credits, approximately 20 hours of TTS, 83 hours of STT, 63 hours of STT translation, and 8 to 17 hours of speech-to-speech translation. Includes 1,000 instant voice clones and commercial use.
M
$340 per month
9 million included credits, approximately 200 hours of TTS, 833 hours of STT, 625 hours of STT translation, and 83 to 167 hours of speech-to-speech translation. Includes 1,000 instant voice clones, 5 Pro Voice Clones, and commercial use.
L
$1,615 per month
45 million included credits, approximately 1,000 hours of TTS, 4,167 hours of STT, 3,125 hours of STT translation, and 417 to 833 hours of speech-to-speech translation. Includes 1,000 instant voice clones, 20 Pro Voice Clones, and commercial use.
Enterprise
Custom pricing
Unlimited credits, custom concurrency, unlimited instant and Pro Voice Clones, commercial use, and enterprise-level arrangements.

Social Listening