Skip to main content
Browse documentation

Creative Tools

Local AI Tools

Browser-side AI models, text-to-speech, music generation, cache, and unload.

How local AI works

  • Local AI runs in the browser when the required APIs and hardware support are available.

  • The first model download can be large and is cached locally afterward.

  • The Local AI status pill shows loading, active, ready, and error states.

Generation tools

  • Use Text to Speech to synthesize speech from text and insert linked audio.

  • Choose supported TTS engines, languages, voices, expressive tags, and speed where available.

  • Use Music Generation with prompts that describe genre, mood, tempo, and instrumentation.

  • Save generated music to the media library or insert it into the timeline.

Cache and runtime control

  • Use Local AI Model Cache to inspect or clear cached downloads.

  • Use Unload Local Models to release resident runtimes.

  • Cancel long-running jobs when you no longer need the result.

Documentation feedback

Was this page helpful?

Your answer helps us decide which guides need clearer steps or examples.