AI Studio and local models
AI Studio keeps image, voice, and audio processing in the same Mac workspace. Tools may run with local models or License-gated server services; the UI shows the required model and current status.
Tools
| Tool | Purpose |
|---|---|
| AI Images | Text-to-image, reference editing, history, retry, and local gallery |
| Model Warehouse | Download and inspect local models |
| TTS | Synthesize speech with presets or an authorized cloned voice |
| ASR | Transcribe media with segments or word timing |
| Audio Restore | Denoise, restore, and upscale audio |
| Source Separation | Split vocals, drums, bass, and other stems |
| Music | Generate local background music assets |
Generated images are stored in the local gallery. ASR can export SRT, VTT, or TXT and optionally label speakers. Voice cloning works best with clean reference audio and must only be used with permission.
Local models can be large. Check available disk space, keep a stable connection during download, and avoid force-quitting during initialization. Review generated or transformed results before adding them to a project.
When you explicitly select a server AI feature, the required input is sent to that service for the requested task. See Data, privacy, and backups.