Garnet is a cutting-edge solution designed to meet the needs of regulated industries, such as government and eHealth, which have stringent data privacy requirements.

Leveraging confidential computing, Garnet enables secure interactions with large language models (LLMs) like Llama and ChatGPT as well as your specific small language models. This product ensures that sensitive information remains protected throughout the data interaction process.

By using advanced security measures, including confidential servers and vector databases, Garnet keeps all sensitive information confidential while allowing users to benefit from the power of LLMs.

The demand for utilizing LLMs like ChatGPT is high, but regulatory constraints make it difficult.

Discussions with enclaive.io customers, particularly in government and eHealth, highlight a significant need - they want to use these technologies but aren't allowed to due to data privacy concerns.

Garnet targets these sectors by taking documents, vectorizing them, pre-filtering, generating prompts, and pseudonymizing them before contacting ChatGPT or another LLM.

This process ensures that no sensitive company data is exposed to the LLM, maintaining GDPR compliance. Technically, it uses a Qdrant vector database and runs entirely in a confidential enclaive cloud.