If you're using Claude, Cursor, or Windsurf with MCP servers, you've probably hit this wall: you find a great server, configure it locally, and then spend the next hour explaining to your colleagues how to do the same. JSON configs in Slack messages. "Did you add the right environment variable?" Version mismatches between machines. Someone installs the wrong server. IT has no idea what AI tools are running on company infrastructure.
MCPNest Enterprise solves this.
What Is MCPNest Enterprise?
MCPNest (mcpnest.io) is the marketplace for MCP servers — 7,561+ indexed, one-click install for Claude Desktop, Cursor, and Windsurf. Enterprise is the layer built on top of it for teams that need to manage MCP servers at scale, with governance, auditability, and control.
Think of it as the npm private registry for your team's MCP servers. Or Docker Hub with access control. The same idea: your team gets a curated, approved set of tools — without anyone editing JSON files manually.
The Core: Private Workspace
Every Enterprise workspace is a private environment that only your team can access. You create it, invite your colleagues, and from that point forward you have a single source of truth for which MCP servers your team uses.
No more "which version are you on?" No more "where's the config file for that?"
What lives inside a workspace:
- Private server registry — add any MCP server (remote endpoint or local config) to your workspace. It stays internal and is never exposed publicly.
- One registry URL — a single authenticated endpoint that your team points their Claude Desktop, Cursor, or Windsurf config to. One change on the server side, everyone gets it.
- Version control per server — publish versions (v1.0.0, v1.1.0), track what changed, roll back if something breaks.
- Audit log — every action is logged: who added a server, who published a version, who invited a member, when.
- Health monitoring — the workspace runs health checks on every remote server. You know immediately when something is degraded or down, before your team starts filing bug reports.
Team Management
Enterprise workspaces have three roles: Owner, Admin, and Member.
The owner has full control. Admins can add servers, publish versions, and invite members. Members can install and use what's been approved for them — nothing more.
This matters for enterprise adoption. The platform engineer sets up the workspace and curates the approved servers. Developers self-serve within those guardrails. IT can see the audit log. Everyone wins.
Inviting a colleague is one email address. They get access immediately if they have an MCPNest account. Remove them just as easily — access is revoked instantly.
The Registry Endpoint
This is the most important technical feature.
Every workspace gets a unique authenticated URL:
https://mcpnest.io/api/registry/YOUR_WORKSPACE_TOKEN
Your team adds this to their Claude Desktop config once:
{
"mcpServers": {
"mcpnest-registry": {
"command": "npx",
"args": ["mcpnest-registry-client", "--url", "https://mcpnest.io/api/registry/YOUR_TOKEN"]
}
}
}
From that point forward, every server you add to the workspace is automatically available to everyone on the team. No individual config changes. No manual distribution. The registry is the single source of truth, and it updates centrally.
Version Control
MCP servers evolve. Endpoints change. Configs need updating. Without version control, you're back to the Slack message problem — "hey everyone, update your config to use the new endpoint."
With MCPNest Enterprise, you publish a new version to the workspace. Set the version number, write a changelog entry, update the install config if needed. Everyone on the team gets the update automatically through the registry endpoint.
If the new version breaks something, you can see exactly what changed and when. The audit log has the full history.
Health Monitoring
Every workspace has a live health dashboard showing which servers are Healthy, Degraded, or Down.
Health checks run automatically every 60 minutes against every remote endpoint in your workspace. If a server goes degraded — timeout, authentication failure, unexpected response — you see it in the dashboard immediately. Before your developers start wondering why Claude isn't responding to tool calls.
You can also trigger a manual health check at any time.
Audit Log
Every action in a workspace is logged with the user, the action type, the timestamp, and any relevant details.
- Workspace created
- Server added
- Version published
- Member invited
- Member removed
- Health check run
This isn't just for debugging. It's for compliance. Enterprise teams need to know what AI tools are running on their infrastructure and who approved them. The audit log gives you that.
The Customisable Dashboard
The workspace overview is designed to show what matters to your team, not a generic set of metrics.
You get five KPI cards — Servers, Members, Installs, Events, Remote — and a grid of widgets you can toggle on and off: the server list, health overview, activity feed, registry endpoint, team members, and quick actions to MCPNest tools.
Your workspace, your layout. Preferences are saved per workspace.
The dashboard also supports light and dark mode — persisted in your browser across all workspace pages.
What's Coming
Enterprise is the foundation. What's being built on top of it:
MCP Gateway (launching June 2026) — one URL per workspace, with routing, authentication, and logging for every tool call. The difference between the current registry and the Gateway is observability: with the Gateway, you see every tool invocation, every parameter, every response. Full auditability at the protocol level.
Invite emails via Resend — when you invite a team member, they receive a proper onboarding email.
GitHub sync — link a server to a GitHub repository and the workspace automatically picks up new releases.
Why This Matters
MCP is the standard that's emerging for how AI agents connect to tools. Every serious AI-powered development team will eventually need to manage a set of MCP servers — which ones are approved, which versions are running, who has access to what.
Right now, most teams are doing this manually. JSON configs, documentation wikis, Slack messages. That doesn't scale past five people.
MCPNest Enterprise is the infrastructure layer that makes MCP adoption manageable at team scale. The same way npm private registries made Node.js dependency management manageable for enterprise teams.
Get Started
MCPNest Enterprise is available at mcpnest.io
Create a workspace, add your first server, and share the registry URL with your team.
If you're building a team AI tooling stack and want to talk, reach out at malasartes@mcpnest.io.
Ricardo Rodrigues is the founder of MCPNest and a Platform Engineer at BCP. MCPNest is built in Porto, Portugal.
Top comments (0)