Hi there,

Cloaked AI is launched! We've been looking forward to this day for several months now, and I'm excited to share the full product with you. We believe Cloaked AI will unlock a lot of potential for companies who are eager to benefit from gen AI but realize the importance of protecting sensitive data.

The Future of AI is Privacy

Cloaked AI is an SDK that protects vector embeddings with data-in-use encryption and is a major breakthrough for companies building AI into their applications using vector databases that hold confidential information. Cloaked AI is...

• Built for easy integration and quick adoption with just a few lines of code
• Integrates with existing vector database or FAISS indices
• Access to data is restricted to permissioned systems with no backdoor access by hosting providers or database admins
• Easy to add on advanced key management and BYOK functionality with IronCore Labs SaaS Shield
• Prevents embedding inversion and membership inference attacks
• Protects associated sensitive data/metadata as well as vectors
• Highly performant with negligible impact on query latency

Many AI Techniques, One Solution

Vector databases are overflowing with sensitive data. But many people might not be familiar with all the of the types of AI systems that use vector databases or what sensitive data pollutes these databases.

These are the top six use cases of both Cloaked AI and vector databases.

1. Recommendation systems 
2. Retrieval augmented generation (RAG)
3. Biometric systems
4. Anomaly and fraud detection
5. Similar image search
6. Semantic text search

We put together a list of the data and security risks associated with each of these systems and what you can do about it.

How easy is it?

Below are two examples of how to encrypt a vector before saving it when using Cloaked AI together with SaaS Shield to handle key management.

Python example

# pypi: ironcore-alloy
plaintext = PlaintextVector([1.2, -1.23, 3.24, 2.37], "contacts", "conversation-sentiment")
metadata = AlloyMetadata.new_simple("tenant-123")
encrypted = await sdk.vector().encrypt(plaintext, metadata)

# Store off encrypted_vector and paired_icl_info in your chosen vector database

Java example

// maven: ironcore-alloy
PlaintextVector plaintext = new PlaintextVector(List.of(1.2f, -1.23f, 3.24f, 2.37f), "contacts",
    "conversation-sentiment");
AlloyMetadata metadata = AlloyMetadata.Companion.newSimple("tenant-123");
EncryptedVector encrypted = sdk.vector().encrypt(plaintext, metadata, null);

// Store off encryptedVector and pairedIclInfo in your chosen vector database

If you're ready to get started protecting vectors, check out our pricing page and the commercial license. If you want to offer your customers BYOK or have us orchestrate key management, check out SaaS Shield, which integrates nicely with Cloaked AI.

Let me know if you have any questions! We're doing a lot at IronCore Labs surrounding AI data protection, and I'd love to hear from you if you have an issue that's not being addressed.

Patrick Walsh
CEO, IronCore