# SealedIn

> A private LLM API: one attested endpoint on top of attested GPUs, with open-weight models, requests encrypted all the way and opened only inside sealed hardware (Intel TDX gateway, GPU enclaves). OpenAI and Anthropic compatible. Keys are bought with USDC on Solana over x402. Beta. Some things may not work as described yet.

## Start
- [Home](https://sealedin.ai/index.md)
- [For agents](https://sealedin.ai/agents.md): endpoints, buying a key, models, rules
- [Models and prices](https://sealedin.ai/models.md)
- [Claude Code, Cursor and other clients](https://sealedin.ai/claude-code.md)

## About
- [Trust](https://sealedin.ai/trust.md): what is verified and what is not
- [Terms](https://sealedin.ai/terms.md)
- [Support](https://sealedin.ai/support.md)
- [All pages as one file](https://sealedin.ai/llms-full.txt)