ExploitDB RAG
LiveExploitDB RAG is a Retrieval-Augmented Generation assistant for searching and analyzing exploits from ExploitDB. It combines intent-aware retrieval, hybrid semantic search, conversational context, and response validation over 46,000+ indexed entries.
ExploitDB RAG features & capabilities:
- Search 46K+ ExploitDB entries by CVE, software name, vulnerability type, or natural language
- Intent classification that routes queries to exact CVE match or semantic similarity search
- Hybrid retrieval with ChromaDB and HuggingFace embeddings
- GPT-4o-mini answers grounded in retrieved exploit documents
- Multi-turn conversation memory for follow-up security queries
- Hallucination validation against source documents and exploit code customization
A practical pentesting assistant that turns ExploitDB into grounded, conversational exploit intelligence.
Stack used
FastAPIPythonReactTypeScriptLangChainHuggingFaceOpenAI