LLMTor

LLMTor offers a robust privacy layer for interacting with large language models, ensuring your anonymity. Key features include: * Identity-prompt cryptographic decoupling * Tor network integration for enhanced privacy * Blind RSA signatures for unlinkable tokens * No user accounts or usage logs kept * Desktop application for secure access This tool is designed to provide cryptographic guarantees that neither the upstream language model provider (like OpenAI or Google) nor the LLMTor operator can link your identity to your specific prompts. By leveraging blind RSA signatures combined with the Tor network, it creates a secure channel where your identity is completely separated from your interactions. The process involves purchasing credits that are linked to your identity for billing, followed by a blind token issuance where a server cryptographically signs a token without ever seeing its contents. You then use this unblinded, unlinkable token to submit prompts via Tor. The dedicated desktop application, available for macOS, Windows, and Linux, handles all the complex privacy mechanisms automatically, including Tor routing and blind token signing. This ensures a seamless and secure experience, making it ideal for privacy enthusiasts, researchers, and journalists who need to run sensitive queries without leaving a trace. Developers can also build upon its open protocol, and organizations can provide employees with confidential access to commercial language models without identity leakage or usage logs.