Inference
Verified before every call.
An OpenAI-compatible API for open-weight models inside trusted execution environments (TEEs). The SDK verifies the enclave's remote attestation, then encrypts each request to it.
Join waitlistFeatures
Your OpenAI code, on private inference.
Swap the client for the Thinkr SDK, and keep the rest of your code.
OpenAI-compatible
The request and response formats of the OpenAI API, streaming included, so existing code moves over with a new client.
Verified before each call
The SDK verifies the enclave's remote attestation against its published measurement, and sends nothing if the verification fails.
Encrypted to the enclave
Each request is encrypted with Hybrid Public Key Encryption (HPKE) to a key that only the enclave holds, so every server on the way sees ciphertext.
Tool calling
Models call your functions and APIs, so agents and workflows run on private inference too.
Vision
Send images and screenshots to the models that read them.
Structured outputs
Ask for JSON that follows your schema, ready for the next step of your code.
Specs
What protects each request.
- API
- OpenAI-compatible
- Encryption
- HPKE, RFC 9180
- Cipher
- AES-256-GCM
- CPU
- AMD SEV-SNP · Intel TDX
- GPU
- NVIDIA, confidential computing
- Verification
- Remote attestation
- Verified by
- The SDK
- Failed verification
- Nothing sent
- Thinkr sees
- Ciphertext
Models
Open-weight models, ready to call.
Every model of the private lane, with its context window and what it can do.
| Model | Context | Can |
|---|---|---|
| GPT-OSS 120B | 131K | Thinking |
| DeepSeek V4.1 Flash | 1M | Images · Thinking |
| Kimi K3 | 256K | Images · Thinking |
| GLM-5.3 | 1M | Thinking |
| Gemma 4 31B | 256K | Images · Thinking |
| Llama 3.3 70B | 128K | Text |