llama.cpp launches llama.app: one-line local LLM install, zero telemetry

ngxson · x · 2026-09-09

llama.cpp launched llama.app (123.9K GitHub stars): frontier AI entirely on your machine with no API keys, no telemetry, no limits. One-line install (curl -LsS https://llama.app/install.sh | sh), one binary spanning Apple Silicon to H100/B200, plus a pi-llama plugin that lets the Pi coding agent auto-discover local models. The post mocks OpenAI for 'highlighting the importance of local, privacy-first inference.'

Original post →

More from coding & agent

coding & agent channel →