Both offer stronger privacy than standard inference. TEE(Trusted Execution Environment) processes your prompt inside a secure hardware enclave, so privacy is enforced by hardware. E2EE (End-to-end encryption) goes further by encrypting your prompt on your device before it leaves, so it stays encrypted until decrypted inside the enclave. Because E2EE keeps prompts encrypted end to end, some features such as web search and memory are disabled in E2EE mode, since they would require decrypting the prompt outside the enclave. Use E2EE for the strongest privacy, or TEE for hardware-verified privacy with broader feature support.
Read the E2EE announcement. For the full picture of Venice's privacy architecture and modes, see venice.ai/privacy.