mcp-hub
Click on "Deploy Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type
@followed by the MCP server name and your instructions, e.g., "@mcp-hublist all available widgets in the hub"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
mcp-hub
Interactive micro-tools for humans and AI agents, served from
mcp.gholl.com.
For humans — a login-free playground of fast, single-purpose widgets.
For AI agents — the same widgets exposed as MCP tools with sandboxed iframe apps (MCP Apps).
Routes
Path | Purpose |
| Tool gallery + playground |
| SEO landing page for a single tool |
| Sandboxed widget view (iframe / MCP Apps) |
| MCP endpoint (Streamable HTTP / JSON-RPC) |
| Agent discovery document |
Related MCP server: Enterprise MCP Gateway and Tool Registry
Stack
Vite · React 19 · TypeScript · Tailwind CSS v4 · @mcp-ui/* · Cloudflare Workers + static assets.
How the UI cards are delivered
Each MCP tool returns a small text + structuredContent result and points at a
UI resource (_meta.ui.resourceUri = ui://gholl/<tool>). Hosts fetch the card
HTML via resources/read; the worker returns the built single-file app with a
window.__GHOLL__ bootstrap injected so it renders that specific widget and
speaks the MCP Apps protocol over postMessage. The same widgets are also
available as regular pages (/tools/:id) and iframes (/embed/:id).
Development
pnpm install
pnpm dev # site at http://localhost:5173
pnpm build # production build to dist/
pnpm lint
pnpm typecheck
pnpm dev:worker # build + serve dist and the Worker locally (wrangler dev)Deployment (Cloudflare)
The project deploys as a Worker with static assets (wrangler.jsonc):
main→worker/index.tshandles/mcp,/mcp/sseand/.well-known/mcp.json.assets.directory→./dist(the Vite build), withnot_found_handling: "single-page-application"for client-side routing.
Build command pnpm build, deploy command pnpm exec wrangler deploy
(or click Deploy in the Cloudflare dashboard with the repo connected).
Bind the custom domain mcp.gholl.com to the Worker.
pnpm deploy # pnpm build && wrangler deployllms.txt, sitemap.xml and robots.txt are generated at build time from the
tool registry in shared/tools.ts. /.well-known/mcp.json is served by the
Worker from the same registry, so it always stays in sync.
See docs/integrations.md for connecting the server to
Claude, Cursor, VS Code, LibreChat, Dify, FastGPT and more. If you are building
an MCP host, docs/host-integration.md is the
implementation checklist (transport requirements, the resources/read UI flow,
and the full MCP Apps postMessage method table). Server metadata for directory
submissions lives in server.json.
Adding a tool
Add a
ToolMetaentry inshared/tools.ts(single source of truth).Build the widget component in
src/widgets/<id>/and register it insrc/widgets/registry.ts.Done — the catalog, landing page, embed route, MCP tool list and discovery document all pick it up automatically.
This server cannot be deployed
Maintenance
Related MCP Connectors
Give any MCP-compatible AI assistant a builder for live, hosted web tools and workflows.
The OpenRouter for tools. One MCP connection gives any AI agent 254 hosted tools, pay per call.
Zero-setup MCP gateway securely connecting AI to your tools with authentication and workflows
Pay-per-use tool marketplace for AI agents. Search, price-check, and call APIs via MCP.
Related MCP Servers
- AlicenseNot gradedqualityCmaintenanceExposes identity, tools, workflows, guardrails, and evaluation as MCP tools — so any AI agent can read and write your ecosystem programmatically.14 npmMIT
- AlicenseNot gradedqualityDmaintenanceEnables AI agents to discover and execute tools via a secure MCP server with JWT authentication, RBAC, rate limiting, and audit logging.1MIT
- AlicenseNot gradedqualityDmaintenanceEnables AI agents to access a unified catalog of tools from various APIs (OpenAPI, GraphQL, MCP, Google Discovery) through the MCP protocol.173 npmMIT
- AlicenseNot gradedqualityCmaintenanceEnables AI agents to securely discover, invoke, and manage tools through a hardened MCP endpoint with protections like injection detection, circuit breakers, retry backoff, response caching, context-window limiting, and state snapshots.4 npmMIT