✦
Summary
Privatemode AI is a confidential AI service operated by Germany-based Edgeless Systems that prioritizes data privacy while delivering generative AI capabilities. It ensures user data remains encrypted throughout its lifecycle, from input to processing and output, using confidential computing technology, so that even the service provider cannot access prompt or response content.
Key Features
End-to-end encryption of prompts and responses during processing
Confidential computing with hardware-backed memory protection
Zero-trust architecture so the provider cannot access user data
OpenAI- and Anthropic-compatible API endpoints
SDK and local proxy for client-verified encryption and remote attestation
Support for open-weights AI models
Hosted in EU data centers with no training on user data
Browser chat app for everyday use
Free tier with 5 million initial chat tokens plus 1 million per month
Open-source components available on GitHub
About Privatemode AI
The service is available as a browser chat app and a developer-focused API with OpenAI- and Anthropic-compatible endpoints, plus an SDK and a local proxy for client-verified end-to-end encryption with remote attestation. It supports open-weights models, runs inference in EU data centers, and does not use prompts or responses for training. Hardware-backed cryptographic certificates verify the integrity of the backend before protected data is sent.
Privatemode AI offers a no-card free tier with an initial 5 million chat tokens plus 1 million tokens each month, with usage-based pricing beyond that. It is aimed at individuals and organizations handling sensitive information, such as healthcare providers, financial institutions and legal firms, and at developers who need to add secure AI features to their applications quickly.