Netbird MCP Server
MCP server that provides read-only access to Netbird network configuration and status, enabling LLMs to analyze network peers, groups, policies, posture checks, networks, nameservers, and port allocations.
Get this MCP server
MCP server that provides read-only access to Netbird network configuration and status, enabling LLMs to analyze network peers, groups, policies, posture checks, networks, nameservers, and port allocations.
Installation
From Source
git clone https://github.com/aantti/mcp-netbird
cd mcp-netbird && make install
Go Install
go install github.com/aantti/mcp-netbird/cmd/mcp-netbird@latest
Smithery
npx -y @smithery/cli install @aantti/mcp-netbird --client claude
Docker
docker build -t mcp-netbird-sse:v1 -f Dockerfile.sse .
docker run --name mcp-netbird -p 8001:8001 -e NETBIRD_API_TOKEN=<your-api-token> mcp-netbird-sse:v1
Configuration
Codeium Windsurf
{
"mcpServers": {
"netbird": {
"command": "mcp-netbird",
"args": [],
"env": {
"NETBIRD_API_TOKEN": "<your-api-token>"
}
}
}
}
Available Tools
| Tool | Description |
|---|---|
list_netbird_peers |
List all network peers |
list_netbird_port_allocations |
List all inbound ports for a specific peer ID |
list_netbird_groups |
List all groups |
list_netbird_policies |
List all policies |
list_netbird_posture_checks |
List all posture checks |
list_netbird_networks |
List all networks |
list_netbird_nameservers |
List all nameserver groups |
Features
- Uses Netbird API to access configuration and status
- Configurable API endpoint
- Secure token-based authentication for Netbird API
- 1:1 mapping of selected Netbird API resources to read-only tools
Environment Variables
Required
NETBIRD_API_TOKEN- Your Netbird API token
Optional
NETBIRD_HOST- Netbird API host (defaults to api.netbird.io)
Usage Examples
Can you explain my Netbird peers, groups and policies to me?
Notes
The project is still in development. Based on the Grafana MCP server from Grafana Labs and uses MCP Go from Mark III Labs. Requires a Netbird API token from the management console. Can be run via Docker, ToolHive, or manually for development.