Teti is a private AI assistant that operates on a subscription model rather than monetizing user data, ensuring conversations are never used for training or shared with third parties. It features persistent memory that remembers user projects and style across sessions while incorporating Lucid, an open-source cognitive-health layer that adapts responses to strengthen reasoning and reduce over-reliance. The platform is powered by proprietary Teti series models built on open-source foundations, with governance overseen by an independent ethics committee.
Funding
Funding not disclosed
Founders
Product
Problem
Most AI assistants train their models on user conversations, turning private thoughts, work, and ideas into data that fuels systems serving other people. This creates an unacceptable trade-off where users pay for a service but also become the product, with no guarantee that their personal information remains confidential.
Solution
Teti provides a private AI assistant that lives off user subscriptions rather than data, ensuring conversations are never used for training or shared with anyone else. The assistant features persistent memory that remembers user projects, style, and work in progress across sessions while keeping all information strictly between the user and Teti. It also incorporates Lucid, an open-source cognitive-health layer that adapts how Teti responds to strengthen reasoning, memory, and original thought rather than replacing them. The platform is powered by proprietary Teti series models built on open-source foundations, allowing independent researchers to verify model behavior and ensuring transparency in how the system operates.
Target Audience
Primary users are privacy-conscious individuals and professionals who want an AI assistant that protects their data and cognitive independence, including those who handle sensitive work or personal information and seek an alternative to data-harvesting AI platforms.
Features
- Persistent memory system that retains user projects, style, and work-in-progress across sessions without feeding data into external models
- Lucid cognitive-health layer that detects risk of over-reliance and suggests pauses to protect user reasoning and creativity
- Proprietary Teti series models built on open-source foundations for independent verifiability and transparency
- Public Ethical Charter with six sealed articles, overseen by an independent Ethics Committee with veto power
- No training on user conversations, no behavior tracking, and no data selling as core architectural principles
- Published system prompt that shapes Teti's behavior, available for public review