Breeze Blue
Blog

Ideas, research, and updates on building with voice

BreezeRT: Serving Ultra-Low-Latency TTS

BreezeRT: Serving Ultra-Low-Latency TTS

How BreezeRT reduces first-audio latency for Breeze TTS 2 with stage scheduling, progressive chunking, CUDA Graphs, and playback deadlines.

Breeze TTS 2 Is #1 on the Artificial Analysis Open-Weights TTS Leaderboard

Breeze TTS 2 Is #1 on the Artificial Analysis Open-Weights TTS Leaderboard

Breeze TTS 2 is the #1 open-weights text-to-speech model in the Artificial Analysis Provider Voice Arena, with a Quality Elo score of 1215 and a 90-point lead over the second-ranked open-weights model, based on blind human listening tests.

Introducing Breeze TTS 2

Introducing Breeze TTS 2

Design any voice. Direct every performance. In real time. Breeze TTS 2 brings natural-language voice design, voice direction, and low-latency streaming generation together in a single model built for interactive experiences.

How to Create Professional YouTube Voice-Overs with AI

How to Create Professional YouTube Voice-Overs with AI

A practical guide to AI narration for YouTube: choosing a voice for your channel, directing tone and pacing in plain language, keeping one consistent voice across uploads, and regenerating a line instead of re-recording it.

Introducing Breeze CLI: Voice AI That Lives in Your Terminal

Introducing Breeze CLI: Voice AI That Lives in Your Terminal

Breeze CLI brings Breeze voice generation to the terminal: hear a voice in seconds, turn any command into a Developer API request, and give AI agents a JSON-native way to create audio and narrated video.

Breeze TTS: Bringing Instruction Following to Speech Generation

Breeze TTS: Bringing Instruction Following to Speech Generation

Breeze TTS is BreezeBlue's instruction-following speech model for controllable AI voice design, voice direction, and expressive generation.