news · 2026-09-08

News: Local-LLM Companion Apps Mature

A wave of companion apps now runs small LLMs entirely on-device (phone or mini-PC), removing the "my doll sends my conversations to a cloud" problem that has rightly worried this community. Expression/timing engines still often run in cloud tiers, so the privacy picture is app-dependent.

Status: verified across multiple vendor releases. Our take: the privacy-conscious setup remains local LLM + Intiface on a mini-PC; but on-device apps are closing the gap fast, and they matter for the AI-head products' roadmap.