Audio & Voice

Turn English speech into punctuated text with word timestamps using an NVIDIA open ASR model.

Generate or extend short music and sound effects from a prompt.

Generate natural speech locally or in a public demo with a small open-weight TTS model.

Turn text into speech with a small TTS model that can run locally on CPU.

Give the model a video, text, or audio cue and create a scene-aware sound layer or edit.

Upload a noisy recording and get a cleaner speech track without manual audio cleanup.

Fly through music genres and hear the sound change in real time.

Replace a GitHub URL segment to hear an AI-generated podcast about the repository.

Speak to ChatGPT, let it read replies aloud, and use shortcuts to keep the conversation moving without typing.

Turn messy speech into cleaner notes with an offline dictation workflow.

Qwen3-TTS is an open multilingual speech family for voice cloning, expressive generation, streaming, and instruction-based delivery.

Turn a script or short voice sample into expressive, editable audio.

Turn lyrics into a song or test a cover and stem workflow with a small free allowance.

Record, clean, edit, and transcribe podcast audio in a browser without installing a desktop editor.

Paste song links and hear a quick mashup or DJ set in the browser.

A controllable text-to-speech and voice-production studio for turning scripts into natural-sounding audio in many languages.

Create voice content or check and protect synthetic media through one voice and security platform.

Turn a style idea into a song draft you can shape.