This tool is a security-hardened fork of the NotebookLM Model Context Protocol (MCP) server, specifically designed to protect sensitive data when AI agents interact with Google NotebookLM. It addresses the inherent risks associated with handling browser sessions, stored cookies/tokens, and proprietary query history by integrating 14 robust security layers. Key protections include post-quantum encryption, secrets scanning, certificate pinning, memory scrubbing, and tamper-evident audit logging, ensuring enterprise-grade security for zero-hallucination answers from your NotebookLM documents across Linux, macOS, and Windows platforms.
Key Features
01Tamper-Evident Audit Logging (with hash chains)
020 GitHub stars
03Post-Quantum Encryption (ML-KEM-768 + ChaCha20-Poly1305 hybrid)
04Memory Scrubbing (zeros sensitive data after use)
05Secrets Scanning (detects 30+ credential patterns)
06Certificate Pinning (blocks MITM attacks on Google connections)
Use Cases
01Securely integrating Google NotebookLM with AI agents (e.g., Claude, Codex, Cursor).
02Conducting autonomous research with zero-hallucination answers from curated documents.
03Protecting sensitive browser sessions, credentials, and query history during AI-driven information retrieval.
