pst

Install or repair the local mlx-audio tooling that pst's text-to-speech depends on (uv tool install + spaCy model). Use when the user runs "/pst:setup", is setting up pst for the first time, or when tts fails because mlx-audio / mlx_audio.tts.generate is missing or broken.

Summarize conversation feedback and speak it aloud using the local tts binary (Kokoro TTS via mlx-audio), and control always-on / verbose voice modes. Use when the user runs "/pst:tts", says "read this back to me", "summarize and speak", "text to speech", wants an audio summary, or wants the assistant to keep speaking throughout the session. Trigger on any request to hear something spoken aloud.