Pixelworlds/opnsense-mcp-server
If you are the rightful owner of opnsense-mcp-server and would like to certify it and/or have it hosted online, please leave a comment on the right or send an email to henry@mcphub.com.
The OPNsense MCP Server is a modular Model Context Protocol server that provides access to OPNsense firewall management methods through a type-safe TypeScript interface.
core_manage
Core system functions like backup, reboot, and firmware info.
firewall_manage
Manage firewall rules and aliases.
interfaces_manage
Manage network interfaces.
diagnostics_manage
Perform system diagnostics.
auth_manage
Manage authentication, users, and groups.
OPNsense MCP Server
A modular Model Context Protocol (MCP) server that provides 88 module-based tools giving access to over 2000 OPNsense firewall management methods through a type-safe TypeScript interface.
Features
- Modular Architecture - 88 logical tools (one per module) instead of 2000+ individual tools
- Complete API Coverage - Access to 752 core methods and 1271 plugin methods
- Type-Safe - Full TypeScript support with @richard-stovall/opnsense-typescript-client v0.5.3
- Plugin Support - Optional support for 64 plugin modules
- Smart Organization - Related operations grouped by module for easier discovery
The MCP server acts as a bridge between AI assistants (like Claude Desktop) and your OPNsense firewall, providing secure API access through a modular tool interface.
Usage in Claude Desktop
Usage in Claude Code
Installation
As an MCP Server
This package is designed to be used as an MCP (Model Context Protocol) server with AI assistants like Claude Desktop, Cursor, or other MCP-compatible clients.
Prerequisites
- Node.js 18 or higher
- An OPNsense firewall with API access enabled
- API key and secret from your OPNsense installation
Install from npm
npm install -g @richard-stovall/opnsense-mcp-server
Usage as an MCP Server
Claude Desktop Configuration
Add the following to your Claude Desktop configuration file:
macOS: ~/Library/Application Support/Claude/claude_desktop_config.json
Windows: %APPDATA%\Claude\claude_desktop_config.json
{
"mcpServers": {
"opnsense": {
"command": "npx",
"args": ["-y", "@richard-stovall/opnsense-mcp-server"],
"env": {
"OPNSENSE_URL": "https://192.168.1.1",
"OPNSENSE_API_KEY": "your-api-key",
"OPNSENSE_API_SECRET": "your-api-secret",
"OPNSENSE_VERIFY_SSL": "false"
}
}
}
}
Alternative Configuration Methods
Using Command Line Arguments:
{
"mcpServers": {
"opnsense": {
"command": "node",
"args": [
"/path/to/opnsense-mcp-server/index.js",
"--url",
"https://YOUR-OPNSENSE-IP",
"--api-key",
"YOUR-API-KEY",
"--api-secret",
"YOUR-API-SECRET",
"--no-verify-ssl"
]
}
}
}
Enable Plugin Tools:
To include all 64 plugin module tools, add "--plugins"
to the args or set "INCLUDE_PLUGINS": "true"
in env.
Testing the Setup
Once configured, you can test the connection by asking Claude:
- "What MCP tools are available?"
- "Use core_manage to get the system status"
- "Use firewall_manage to search for all aliases"
- "Use interfaces_manage to list all network interfaces"
Troubleshooting Claude Desktop Setup
Connection Issues:
- Verify your OPNsense API is enabled
- Check that the API key has appropriate permissions
- Ensure the IP/hostname is accessible from your machine
- For self-signed certificates, use
--no-verify-ssl
or set"OPNSENSE_VERIFY_SSL": "false"
View Server Logs: Check Claude Desktop logs for any error messages from the MCP server.
Test Manually: You can test the server manually before using with Claude Desktop:
node /path/to/opnsense-mcp-server/index.js \
--url https://YOUR-OPNSENSE-IP \
--api-key YOUR-API-KEY \
--api-secret YOUR-API-SECRET \
--no-verify-ssl
This should output:
OPNsense MCP server v0.6.0 (modular) started
Core tools: 24 modules
Plugin tools: 64 modules (disabled)
Total available: 24 modules
Cursor Configuration
Add to your Cursor settings (.cursor/mcp.json
in your project or ~/.cursor/mcp.json
globally):
{
"mcpServers": {
"opnsense": {
"command": "npx",
"args": ["-y", "@richard-stovall/opnsense-mcp-server"],
"env": {
"OPNSENSE_URL": "https://192.168.1.1",
"OPNSENSE_API_KEY": "your-api-key",
"OPNSENSE_API_SECRET": "your-api-secret",
"OPNSENSE_VERIFY_SSL": "false"
}
}
}
}
Configuration Options
The server accepts configuration through environment variables:
OPNSENSE_URL
- OPNsense host URL (required)OPNSENSE_API_KEY
- API key for authentication (required)OPNSENSE_API_SECRET
- API secret for authentication (required)INCLUDE_PLUGINS
- Set to "true" to enable 64 plugin module tools (optional)OPNSENSE_VERIFY_SSL
- Set to "false" to disable SSL verification (development only)
How It Works
The modular MCP server provides your AI assistant with 88 module-based tools. Each tool represents an OPNsense module and accepts a method
parameter to specify the operation.
Tool Usage Pattern:
{
"tool": "firewall_manage",
"arguments": {
"method": "aliasSearchItem",
"params": {
"searchPhrase": "web"
}
}
}
Example prompts:
- "Use core_manage to check system status"
- "Use firewall_manage to list all firewall aliases"
- "Use interfaces_manage to get network interface information"
- "Use plugin_nginx_manage to check the web server configuration"
- "Use diagnostics_manage to view the ARP table"
The modular approach makes it easy to discover related functionality - all firewall operations are in firewall_manage
, all VPN operations in their respective modules (openvpn_manage
, ipsec_manage
, wireguard_manage
).
Available Module Tools
Core Modules (24 tools)
Each tool provides access to all methods within that module:
Tool Name | Description | Example Methods |
---|---|---|
core_manage | Core system functions | backupBackups , systemReboot , firmwareInfo |
firewall_manage | Firewall rules & aliases | aliasSearchItem , filterAddRule , natSearchRule |
interfaces_manage | Network interfaces | getInterfaces , vlanAddItem , setInterface |
diagnostics_manage | System diagnostics | interfaceGetArp , systemActivityGetActivity |
auth_manage | Authentication | userSearchUser , groupSearchGroup |
firmware_manage | Firmware updates | check , update , upgrade , changelog |
openvpn_manage | OpenVPN | instancesSearch , instancesAdd , serviceReconfigure |
ipsec_manage | IPsec VPN | tunnelSearchPhase1 , connectionStatus |
wireguard_manage | WireGuard VPN | serverSearchServer , clientSearchClient |
unbound_manage | DNS resolver | hostOverrideSearchItem , serviceReconfigure |
dhcpv4_manage | DHCP server | searchLease , addReservation |
Plugin Modules (64 tools when enabled)
Popular plugin modules:
Tool Name | Description | Example Methods |
---|---|---|
plugin_nginx_manage | Nginx web server | generalGet , upstreamSearchUpstream |
plugin_haproxy_manage | HAProxy load balancer | serverSearchServer , statsGet |
plugin_caddy_manage | Caddy web server | reverseProxySearchDomain , serviceStatus |
plugin_bind_manage | BIND DNS | domainSearchDomain , recordSearchRecord |
plugin_acmeclient_manage | Let's Encrypt | certificatesSearch , certificatesIssue |
Building from Source
If you want to contribute or customize the server:
# Clone the repository
git clone https://github.com/richard-stovall/opnsense-mcp-server.git
cd opnsense-mcp-server
# Install dependencies with Yarn 4.9.2
yarn install
# Build the project
yarn build
# Run locally
yarn start
Development
Development Scripts
yarn generate-tools # Generate tool definitions
yarn build # Build the server
yarn build:all # Generate tools and build
yarn dev # Run with hot reload
yarn type-check # Type check without emitting
yarn start # Start the server
Technology Stack
- Runtime: Node.js with tsx for TypeScript execution
- Package Manager: Yarn 4.9.2 with Plug'n'Play
- Build System: Simple TypeScript compilation to single file
- Language: TypeScript 5.3+
- MCP SDK: @modelcontextprotocol/sdk
- API Client: @richard-stovall/opnsense-typescript-client
- Validation: Zod for schema validation
- Testing: Jest with TypeScript support
API Integration
The server uses the @richard-stovall/opnsense-typescript-client package which provides:
- Complete type safety for all API calls
- Built-in error handling and retries
- Support for all 601 OPNsense API endpoints
- Modern Fetch API based implementation
Example Tool Implementation
const response = await client.system.getStatus();
return {
content: [
{
type: 'text',
text: JSON.stringify(response.data, null, 2),
},
],
};
Contributing
Contributions are welcome! Please feel free to submit a Pull Request.
- Fork the repository
- Create your feature branch (
git checkout -b feature/AmazingFeature
) - Commit your changes (
git commit -m 'Add some AmazingFeature'
) - Push to the branch (
git push origin feature/AmazingFeature
) - Open a Pull Request
License
This project is licensed under the MIT License - see the LICENSE file for details.
Acknowledgments
- Built on the Model Context Protocol by Anthropic
- Powered by @richard-stovall/opnsense-typescript-client
- Inspired by the OPNsense community
Made with love for the OPNsense community