SayGM

About SayGM

SayGM is an LLM gateway that consolidates access to Claude, GPT, Gemini and open models through a single API. The service routes requests through a trusted execution environment (TEE) on Intel TDX, sealing your prompts until they reach the model provider's hardware. For frontier models like Claude and GPT, requests are sealed through the gateway and processed under the provider's terms. For open-weight models, prompts stay encrypted inside the TEE enclave throughout processing.

The platform offers live pricing that is capped at official list prices, with an average reported discount of 37%. It works directly with existing OpenAI and Anthropic SDKs and integrates with tools like Cursor, Cline, Claude Code and the Vercel AI SDK. Users can get an API key in minutes and pay only for what they use with no top-up fees. SayGM Desktop is also available as an AI agent for local file work.

• tee attestation
• llm gateway
• api access
• private inference
• unified api
• pricing comparison
• ai models

FAQ

How does SayGM keep my prompts private?

For frontier models (Claude, GPT, Gemini), the gateway seals your request until it reaches the model provider's hardware under their API terms. For open-weight models, your prompt stays encrypted inside the TEE enclave, decrypted only within the enclave that runs the model. The gateway runs in an Intel TDX enclave whose measurements are hardware-signed by the chip itself, allowing anyone to verify the exact code running.

Can I use SayGM with my existing code?

Yes. SayGM works with your existing OpenAI or Anthropic SDK. You point your client at the SayGM gateway with one base URL change, and your existing code keeps working. It integrates natively with tools like Cursor, Cline, Claude Code, and the Vercel AI SDK.

How many models does SayGM support?

SayGM supports 60 models including Claude, GPT, Gemini and open-weight models, with live pricing shown for each. The platform uses provider bidding and eligible offers from miners to determine rates.