CONFSEC is a private AI inference wrapper designed to provide secure and anonymous artificial intelligence inference requests. It ensures that user prompts are never logged, retained, used for training, or accessible to system operators by enforcing stateless computation and verifiable transparency. The system utilizes underlying technologies such as Oblivious HTTP, blind signatures, remote attestation, and hardware-based Trusted Execution Environments (TEEs) to protect private computation from adversarial access.
Developers can interact with the platform using the CONFSEC Web Console, CLI, and dedicated Node.js and Python SDKs. These SDKs function as drop-in replacements for existing OpenAI clients or as direct HTTP clients for lower-level API access, allowing seamless integration without the need to deploy or manage a separate proxy.
