About
We build private AI.
By default, open-weight models answer your chats inside trusted execution environments (TEEs), and not even Thinkr can read them.
Mission
Why we build it.
AI grows more personal, more present, and more capable every year, so keeping our conversations, ideas, memories, and work private is what keeps people at the center. AI should multiply what people can do, never stand in for them. Thinkr builds that protection into the technology itself, with encryption, hardware enclaves, and remote attestation, so no third party can reach what you think: the ground on which people grow, create, and thrive.
The app
What runs today.
Every chat starts in the private lane. Your device first verifies the enclave's remote attestation, and sends nothing if the verification fails. It then encrypts each message with Hybrid Public Key Encryption (HPKE) to a key that only the enclave holds.
The open lane serves Claude, GPT, Gemini, and Grok with zero data retention on every request. Their labels say that the provider can read the chat.
Thinkr loads no trackers and shows no ads, and no chat in either lane is used for training. You can try the private lane free, with no account.