IronCore Labs

Founded 2015

Encrypt vector embeddings and manage per-tenant keys for sensitive data

You can use IronCore Labs to protect sensitive data and AI vector embeddings directly at the application layer. It lets you encrypt embeddings before sending them to vector databases while preserving searchability for nearest-neighbor lookups. The platform provides full lifecycle cryptographic key orchestration, allowing SaaS vendors to offer Bring Your Own Key and Hold Your Own Key options across various KMS providers. It also includes encrypted keyword search and command-line scanning tools to find exposed personal data inside numerical vectors.

What IronCore Labs does

Encrypt vector embeddings before storing them in vector databasesOrchestrate per-tenant encryption keys across AWS, Azure, and GCP KMSExecute nearest-neighbor vector searches over encrypted dataRun encrypted keyword searches on Elasticsearch and OpenSearch indexesScan vector embedding datasets for personally identifiable informationConfigure customer-held encryption keys for SaaS applicationsRotate and revoke cryptographic keys across distributed data storesEncrypts vector embeddings while retaining nearest-neighbor search utilityManages multi-tenant customer-managed encryption keysSupports Bring Your Own Key and Hold Your Own Key workflows across any KMSProvides encrypted keyword search inside Elasticsearch and OpenSearchScans vectors for hidden personally identifiable information using VectorLensOffers open-source SDKs for Rust, Python, Go, Node, Java, and PHPSwaps cryptographic algorithms and key sizes via configuration without code rewrites