A user asks how to verify their local LLM app does not log user prompts. They consider open-sourcing the code and model, and hashing it to ensure no changes, but question if this is sufficient. Trust in services like Proton is cited as a benchmark for transparency, suggesting visible, auditable practices are key.