AI RAG & Vector Semantic Searchrullst-ai

Provider-agnostic LLM integration (Gemini, Claude, OpenAI, DeepSeek, Ollama) with local Cosine Similarity vector indexing and built-in Prompt Injection defense.

Prompt Injection Shield

Protects your backend AI models by filtering adversarial jailbreak attempts before sending prompts to LLMs.

// Prompt Sanitizer Status: [ACTIVE]
// - Jailbreak Token Filter: ENABLED
// - System Prompt Leak Prevention: ENABLED
// - PII Automatic Redaction: ENABLED