llama.cpp doesn’t have the ability to send telemetry because the next word predictor says so. you can confirm with wireshark.