Gradium
Low-latency voice AI API for speech synthesis, transcription, translation and cloning.
Introduction
Gradium offers text-to-speech, speech-to-text, live translation, voice cloning and an offline on-device TTS model behind one API. It targets real-time voice agents, with first audio in about 200 ms, Python and Rust SDKs, and a free tier of 45,000 credits.
- Added on
- Nov 2, 2026
- Website
- gradium.ai
- Pricing model
- Freemium
Product information
What is Gradium?
Gradium is a voice AI platform for developers offering low-latency TTS, STT, speech translation, voice cloning and on-device TTS through a unified API.
How to use Gradium?
Create an account, start on the free tier or pick a plan, and use the Studio or API. Integrate the Python or Rust SDKs or connect via WebSocket streaming, including with LiveKit and Pipecat.
Core features of Gradium
- Streaming text-to-speech
- Speech-to-text with custom vocabulary
- Real-time speech translation
- Voice cloning from about 10 seconds of audio
- Offline on-device TTS on CPU
- Bidirectional WebSocket API
Use cases of Gradium
- 1Build real-time customer support voice agents
- 2Create multilingual voice assistants
- 3Transcribe calls, meetings and interviews
- 4Generate voiceovers for media and advertising
- 5Develop offline voice apps for devices
Frequently asked questions about Gradium
How many credits are in the Gradium free plan?+
The Free plan includes 45,000 credits.
Does Gradium support voice cloning?+
Yes, with instant cloning from about 10 seconds of audio and higher-fidelity Pro Voice Clones.
Which deployment options does Gradium offer?+
Cloud API, dedicated instances, self-hosted, on-premises and cloud marketplaces.
Which category is Gradium in?+
Gradium is listed under AI Voice Generator.