https://caution.co

Demo: AI Inference

This is a demo AI inference application running inside a secure enclave. Your conversation is end-to-end encrypted and only decryptable by you and the enclave.

The main differentiators versus other confidential compute:

verifiability

• true end-to-end encryption.

This agent runs on CPU, GPU support is in active development. TinyLlama 1.1B Chat.

This software is currently in alpha and not production ready.

How can I help you today?

Start a conversation by typing a message below.