# RedPill

> Confidential AI you can verify. An OpenAI-compatible inference API served from a TEE-attested gateway, with a signed receipt on every response. Agents can run inference on it, through Private AI Proxy or a base URL; separately, an agent can manage a RedPill account through the account MCP server.

## Documentation

- [Documentation index](https://docs.redpill.ai/llms.txt): guides, API reference, and how the proofs work.
- [Quickstart](https://docs.redpill.ai/get-started/quickstart): first request and receipt check.
- [Use in agents](https://docs.redpill.ai/get-started/use-in-agents): run inference from Claude Code, Codex, Pi, OpenCode, and more.
- [MCP server](https://docs.redpill.ai/get-started/mcp): let an agent manage a RedPill account.

## Agents

- [Agent guide](https://redpill.ai/agents.md): inference in agents, account management by agents, and the boundary between them.
- [RedPill skill](https://redpill.ai/skills/redpill/SKILL.md): installable Agent Skills instructions for both.
- [Agent Skills discovery index](https://redpill.ai/.well-known/agent-skills/index.json): machine-readable skill metadata and integrity digest.
- [Account MCP server](https://service.redpill.ai/mcp): manage a RedPill account through OAuth.
- [Documentation MCP server](https://docs.redpill.ai/mcp): search public RedPill documentation.

## API

- [OpenAI-compatible API](https://api.redpill.ai/v1): base URL for RedPill inference.
- [Confidential models API](https://tee.redpill.ai/v1): serves confidential models only.
- [Model catalog](https://redpill.ai/models): every model with pricing, context, and confidentiality.

## Product

- [Start building](https://redpill.ai/start-building): the API path and the agent path.
- [Private AI Proxy](https://docs.redpill.ai/get-started/private-ai-proxy): desktop app and CLI that verify the gateway and connect agents.
- [Pricing](https://redpill.ai/pricing)