MCP Tools
llm.port supports governed use of Model Context Protocol (MCP) tools so assistants can safely call external capabilities.
What MCP gives you
- Tool discovery and registration
- Controlled tool invocation
- Consistent governance behavior across tool calls
Typical tool categories
- Web search
- Content retrieval and extraction
- Domain-specific internal tools
Trust and control expectations
- Apply access policies before enabling tools in production
- Review tool outputs in observability workflows
- Restrict tool access by environment and role
Rollout recommendations
- Start with a small trusted tool set
- Validate behavior in staging
- Expand tool scope gradually with audit review
This page focuses on public integration patterns and excludes internal broker implementation details.