AI-Powered Toxic Content and Spam Detection for Drupal

Community-driven Drupal sites currently rely on reactive, manual moderation for spam and toxic content, which is exhausting and difficult to scale. This project introduces a proactive, real-time defense system by building a fully integrated AI Content Moderation module. The solution utilizes a decoupled ContentModerator service that intercepts form submissions on the client side using debounced JavaScript, providing users with instant, WCAG-compliant inline warnings before toxic data ever hits the database. On the backend, it employs a highly performant hybrid approach, checking local administrative keyword lists before securely querying external NLP models (such as Hugging Face) via Drupal's core HTTP client. Deliverables include a production-ready module featuring an intuitive Admin UI for AI provider and threshold configuration, bulletproof server-side validation fallbacks, and comprehensive site builder documentation.