Gradium
Gradium is a voice AI platform for developers and organizations building speech-enabled applications and conversational agents. Its hosted APIs provide streaming text-to-speech, speech-to-text, speech translation, and speech-to-speech, with WebSocket and REST interfaces, real-time turn detection, pronunciation controls, and voice design or cloning features. A browser-based studio lets users try the models and manage voices, while documentation and SDK guides support application development. Use cases include voice agents, customer-service automation, live transcription, multilingual voice experiences, and spoken content generation. Gradium also publishes Phonon, an open-weight text-to-speech model designed for local, offline use in applications and on-device systems. Plans use credits for speech generation, transcription, and translation.
Learn more
Phonon
Phonon is an open-weight text-to-speech (TTS) model and inference runtime for developers adding speech generation to applications that need local, offline audio. It converts text into spoken audio and can stream speech as it is generated. The runtime provides Rust and Python interfaces, a Swift package for iOS and macOS, JavaScript for browsers, and Android libraries; it can also run as a self-hosted server with an OpenAI-compatible API. Depending on the model variant, it supports English, French, German, Spanish, and Portuguese. Phonon is intended for developers integrating speech into mobile, desktop, web, embedded, and server applications, including offline or privacy-sensitive use cases. The 31-million-parameter English model weights use the Apache 2.0 license; 40-million-parameter language-specific and multilingual weights use CC BY-NC 4.0. The runtime code is separately available under Apache 2.0 or MIT.
Learn more
Gemini 3.8 Flash TTS
Gemini 3.8 Flash TTS is Google’s expressive text-to-speech model for creating custom voices, directed performances, and multilingual audio experiences. The model can generate original voices from natural-language prompts by specifying characteristics such as role, accent, tone, pacing, and vocal style across more than 100 languages and dialects. Users can also replicate an authorized voice from a short audio sample, with consent verification, SynthID watermarking, and C2PA credentials supporting responsible voice creation. Gemini 3.8 Flash TTS provides line-by-line performance control, long-form speech generation, two-speaker scene staging, and support for nonverbal cues such as laughs, sighs, gasps, and conversational backchanneling. It is suited to use cases including games, audiobooks, podcasts, dubbing, interactive voice agents, branded audio, and media localization.
Learn more
Gemini 3.8 Flash-Lite TTS
Gemini 3.8 Flash-Lite TTS is Google’s cost-efficient text-to-speech model designed for high-volume dubbing, audio production, and expressive voice-agent applications. The model provides fine-grained control over tone, pacing, expressive nuance, and line-by-line vocal delivery. It supports long-form speech generation while maintaining natural pacing, voice quality, and speaker consistency across extended content. Native two-speaker scene staging enables multi-turn dialogue with distinct voices and natural conversational turn-taking from a single script. Gemini 3.8 Flash-Lite TTS supports more than 100 languages and can add nonverbal cues and backchanneling to create more natural conversational audio. The model is available through the Gemini API and Google AI Studio, with deployment through Google Vids and planned enterprise availability through Gemini Enterprise.
Learn more
