SayGM

0 reviews

SayGM is a private LLM gateway: one API for 60+ frontier and open-weight models, where SayGM can't read your prompts. Every request runs inside a gateway sealed in an Intel TDX enclave, and you can verify the attestation yourself. Confidential models run inside the enclave too, so even the provider can't read them. Guardrails can redact personal data, and prompts are never logged. Prices are per token, capped at each maker's list price, with no subscription. OpenAI, Anthropic and Gemini-compatible endpoints mean existing code keeps working.