Expose your prompt library to Claude — without building an MCP server
PromptHub ships a native MCP server with API-key and OAuth 2.1 authentication, so your Prompts, Chains, Skills, PlugIns, and Agents become governed, RBAC'd tools any MCP client can call.
How PromptHub solves this
Instead of hand-rolling auth, scopes, and a bespoke server, connect via a per-user API key or full OAuth 2.1 — PromptHub's existing role-based access and rate limits apply to every MCP tool call automatically.
What's included
API Key or OAuth 2.1
Connect with a per-user API key or full OAuth 2.1 (PKCE) — works with Claude Code, Claude Desktop, and claude.ai custom connectors.
Governed Tool Calls
Every MCP tool call is scope-checked, RBAC-checked, rate-limited, and audited — the same guard your REST API already uses.
Skills, PlugIns & Agents as Tools
Tools like list_skills, get_agent, and execute_agent expose your governed building blocks directly to any MCP client.