Lucid runs a 244M-parameter clinical voice model entirely on your Mac. No cloud, no wearables. It passively tracks stress, mood, energy, and depression risk from how you speak and surfaces trends before you feel them. Free beta.