Keypost Documentation
Keypost is a policy enforcement layer for MCP (Model Context Protocol) servers. It sits between your AI agents and MCP servers, enforcing access control, rate limits, DLP, and audit logging.
What is Keypost?
When you connect AI agents like Claude or Cursor to MCP servers (GitHub, Slack, databases), those servers expose powerful tools with zero restrictions. Keypost adds a policy layer that lets you:
- Control access - Allow or deny specific tools
- Constrain parameters - Restrict what values can be passed
- Rate limit - Prevent abuse with throttling and quotas
- Scrub sensitive data - DLP patterns block or redact PII, secrets
- Require approval - Human-in-the-loop for sensitive operations
- Audit everything - Full trail of every tool call
How it works
Keypost acts as a transparent proxy. Instead of pointing your agent directly at an MCP server, you point it at Keypost. Keypost evaluates your policies, then forwards allowed requests to the upstream server.
AI Agent → Keypost (policies) → MCP Server
← ←