what is higgs-audio-v3-studio fr?

Higgs Audio v3 Studio is a desktop application for Windows and Linux that lets you generate speech on your own computer using the Higgs Audio v3 text to speech model. It is built with Rust and Tauri for the interface, and runs the actual audio generation through a native C++ engine that uses your NVIDIA graphics card for speed. Because everything runs locally, you do not need to set up Python or send your data to an outside service. The app supports several ways to create audio: normal text to speech, cloning a voice from a short recording, continuing a piece of speech naturally, and generating conversations with multiple different speakers. You can drag and drop a reference voice clip, record one directly from your microphone, or trim an existing clip to the part you want to use, all with waveform previews so you can see what you are working with. It can also read voice notes recorded in apps like Telegram or WhatsApp without needing separate audio conversion software installed. A Speaker Gallery lets you save voices you use often, along with their reference audio, a transcript, and a note, so you can reuse them later without redoing the setup each time. The app also includes an optional automatic transcription feature powered by Whisper, in case you do not already have a written transcript for your reference audio. While audio is generating, you can watch progress update live, including how many tokens the model has produced and how much time has passed. Finished results are tracked per generation mode so you can look back at previous outputs. Generation settings like temperature, speed, pitch, and expressiveness are all adjustable, and finished audio can be exported as a WAV or MP3 file. To use the desktop engine you need an NVIDIA graphics card that supports CUDA 13, along with enough video memory for whichever quantized version of the model you choose to download, ranging from about 8 to 16 gigabytes depending on the version. A Linux build with a matching engine is maintained on a separate branch.