Detects key claims in long messages and summarizes the real point. Uses NLP to find what someone is actually saying vs. what they want you to believe.

This is a small local tool that runs with python3 and prints analysis to stdout. It does not send data over the network or request credentials. Before using: (1) review the included detect.py (already provided) if you have concerns; (2) do not point the --input argument at sensitive system files or secret/config files; (3) be aware the script has minor bugs (e.g., potential division by zero on empty input) but nothing indicating malicious behavior; (4) because the source and homepage are unknown, prefer running it in a restricted environment (local sandbox/VM) if you are cautious.