This document proposes a new approach called ε-MPPI (epsilon-Multikeyword Private Privacy Index) for privacy-preserving search of distributed documents in a network. It aims to quantitatively control privacy leakage for multi-keyword document searches. The key aspects are: (1) Documents are stored encrypted on private servers for added security. (2) Private server indexes are merged and stored on a public cloud for search. (3) Authentication and encryption keys are used to retrieve encrypted results while preserving privacy. The proposed approach was implemented and experiments showed it provides improved performance over existing techniques.