Private AI with the power of cloud computing. In the private lane, open-weight models run inside trusted execution environments (TEEs). Your browser verifies their remote attestation, then encrypts each chat with Hybrid Public Key Encryption (HPKE) before it leaves your device. The open lane reaches frontier models through zero data retention (ZDR) endpoints. The app has no trackers or ads and needs no account. Your chats are never used for training.

For developers, Thinkr Inference is an OpenAI-compatible API whose SDK verifies the enclave's attestation before each call. Thinkr Containers runs any Docker image in a hardware enclave, measured for anyone to verify.

The Thinkr token is the utility token of the app, used for discounts, premium access, and higher usage limits, with a buyback and burn behind it.