TAWA GUARD verifies every input before your AI reads it: prompts, documents, emails, web pages. Allow, flag or block, with the reason. Works with any model. You set the rules, and every decision is logged. Free API key.