Gradium

What is Gradium

Gradium is a voice AI platform that provides ultra-low-latency APIs for text-to-speech, speech-to-text, and voice cloning. It enables developers to generate expressive speech, accurate transcriptions, and high-fidelity voice clones through a single API. The platform is designed for building voice agents and real-time voice applications, with a focus on natural, expressive, and robust performance across multiple languages.

How to use Gradium

1. Sign in to the Gradium platform using your account.
2. Choose a model from the available options: Text-to-Speech, Speech-to-Text, Live Translation, or Voice Cloning.
3. Integrate the API into your application by following the documentation provided.
4. Test the models using the demos available on the website or through the API.
5. Deploy your voice application with the API, leveraging Gradium's low latency and multilingual support.

Features of Gradium

• Text-to-Speech: Turn text into natural-sounding voice with expressive output.
• Speech-to-Text: Transcribe audio into text instantly with high accuracy.
• Live Translation: Convert voice to voice in real time, enabling multilingual communication.
• Voice Cloning: Create instant or professional voice clones with high fidelity.
• On-Device TTS: Offline real-time natural voice generation for on-device applications.
• Ultra-low latency: Optimized for real-time voice interactions.
• Multilingual support: Robust across every language, with support for English, French, Spanish, Portuguese, and German.
• Single API: Access all voice AI models through one unified API.
• Gradbot: Build voice agents in a single prompt.
• Agent Demo: Test voice agents live.

Use Cases of Gradium

• Voice Agents: Build real-time voice agents for customer support, virtual assistants, and more.
• Synthetic Callers: Create realistic synthetic voices for evaluation agents and testing.
• Multilingual Communication: Enable real-time multilingual communication with live translation.
• Enterprise Voice Solutions: Deploy voice agents in enterprise environments with reliable performance.
• Content Creation: Generate natural-sounding voiceovers for videos, podcasts, and other media.
• Accessibility: Provide speech-to-text and text-to-speech capabilities for accessibility applications.

Pricing

Gradium offers a pricing structure with credits. During the public beta, users who provide feedback receive 1 million Gradium credits. Specific pricing plans are available on the Pricing page.

FAQ

What is the public beta? Gradium has released a new Text-to-Speech model in public beta. It handles complex cases natively, such as phone numbers, email addresses, IBAN numbers, and time expressions, without requiring pre-processing or text normalization.

How can I test the new TTS model? Anyone can test it through the API. Users who send feedback during the beta receive 1 million Gradium credits.

What languages are supported? Gradium supports English, French, Spanish, Portuguese, and German, with robust performance across every language.

Is there an on-device option? Yes, Gradium offers On-Device TTS for offline real-time natural voice generation.

How do I get started? Sign in to the Gradium platform, choose a model, and integrate the API using the documentation. You can also try the demos and the Agent Demo to see the capabilities.