Real-time interactive words in chat(open source – MIT) (trustquery.com)

An open-source toolkit called TrustQuery (MIT-licensed) delivers real-time, interactive “words in chat” features for any textarea, enabling input validation, intelligent autocomplete, blocking, and clarifying-question prompts in under five minutes of integration. Its core use cases are preventing PII leaks, catching and correcting malformed queries, and guiding users toward clearer prompts before they send them. The project emphasizes quick client-side augmentation of chat and form UIs so teams can add live protections and UX improvements without heavy backend work. This matters because it addresses common operational and safety pain points in AI deployments: data governance (avoiding accidental data exfiltration), reduced downstream model errors from bad prompts, and higher-quality user inputs that improve model outputs and reduce moderation burden. Technically, TrustQuery appears to be a lightweight, embeddable library that enriches textareas with heuristics and dynamic interactions (autocomplete, blocking rules, clarifying flows) and is extendable; additional modules for AI reliability, trust verification, and business intelligence are planned. Being MIT-licensed makes it production-friendly for commercial projects and encourages community contributions, so teams can adapt rulesets, integrate with existing LLMs or policy engines, and iterate on safeguards quickly.