Searchable Encryption with OPRF for Privacy-Preserving Content Search

Abstract

This paper describes the implementation of searchable encryption using OPRF (Oblivious Pseudorandom Function) with Ristretto255 for the ZKIM file format. The system enables privacy-preserving search over encrypted content without revealing search queries or content to the platform. We present a production-ready implementation that integrates OPRF-based trapdoor generation, encrypted search indexes, bucket policy for side-channel resistance, and epoch-based key rotation. The system achieves privacy-preserving search without data access, enabling users to search their encrypted messages and files while maintaining complete privacy. Performance evaluation demonstrates efficient search operations with sub-second query response times for typical workloads while maintaining strong security guarantees. The implementation integrates with the ZKIM messaging system via `ZKIMMessageIndexingService`, providing real-world deployment data and practical insights into privacy-preserving search at scale.