Model Context Protocol (MCP)
MCP is an open protocol that lets AI agents call external tools and services in a structured, secure way. This section tracks the MCP servers built and operated by LeeP Tech — each with implementation notes and direct endpoint links.
MCP Dataverse
A production-ready Dataverse MCP server focused on reliable agent execution: schema-aware guidance, session continuity across long workflows, dirty-field precision writes, and cleaner multi-agent handoff behavior.
Authentication uses OAuth 2.0 on-behalf-of, so Dataverse operations run as the signed-in user for stronger auditability and safer enterprise usage.
Open MCP Dataverse Server
fastapi-mcp-azure-oauth
RFC-compliant Azure AD OAuth 2.0 router for FastAPI MCP servers. Implements RFC 8414 discovery, RFC 7591 Dynamic Client Registration with automatic Azure AD redirect URI enrolment via Microsoft Graph, RFC 9728 protected resource metadata, and JWT token validation with per-tenant JWKS caching and explicit issuer binding. Built for first-class Copilot Studio compatibility.
Published on PyPI. Requires Python 3.10+ and FastAPI 0.115+.
Open fastapi-mcp-azure-oauth Server