RuleGo-Server exposes MCP (Model Context Protocol) integration. When MCP is enabled, RuleGo-Server automatically registers all components, rule chains, and APIs as MCP tools, allowing AI assistants and other MCP clients to call these tools directly via the MCP protocol. This enables deep integration with automation workflows, API orchestration, data processing, and IoT use cases, with support for a marketplace of components and dynamic loading. You can leverage MCP-enabled tools to orchestrate complex rule chains or invoke API endpoints from AI assistants, enabling seamless cross-system automation.

To use the MCP tools, ensure MCP is enabled in the server configuration (see config.conf under the [mcp] section). You will then access registered MCP tools from your MCP client by querying the server for available tools and invoking them by name or ID as defined by RuleGo-Server. Typical workflows involve discovering tools, then sending calls with appropriate context data to drive automated orchestration across components, rule chains, and APIs exposed by the server. Documentation for the MCP integration is available at rulego-server-mcp pages, and tool registration occurs automatically based on the loaded components, chains, and APIs.

Prerequisites:

• A supported environment with Go installed (1.20+ recommended) or a prebuilt server binary if available.
• Git to clone the repository.

Install steps:

1. Clone the repository: git clone https://github.com/your-org/rulego-rulego-server.git cd rulego-rulego-server

2. Build from source (default build): cd cmd/server go build .

   This produces a binary named 'server' in the cmd/server directory

3. Optionally enable extended components or AI extensions during build via tags as described in the repository (examples):

   Include extension components

   go build -tags with_extend .

   Include AI extension components

   go build -tags with_ai .

   Include multiple libraries at once

   go build -tags "with_extend,with_ai,with_ci,with_iot,with_etl" .

4. Prepare a configuration file (config.conf) as described in the README, including MCP settings under [mcp], and start the server: ./server -c ./config.conf

5. Run in background (optional): nohup ./server -c ./config.conf > server.log 2>&1 &

Notes:

• Ensure data_dir, JWT, and MCP settings in config.conf are configured as needed for your environment.
• If you plan to use MCP tools from an AI assistant, confirm that load_components_as_tool, load_chains_as_tool, and load_apis_as_tool are enabled in the [mcp] section.

Tips and common considerations:

• The MCP feature automatically registers all components, rule chains, and APIs as MCP tools when enabled. This enables AI assistants to call tools through the MCP protocol.
• Ensure data_dir and resource mappings in config.conf are correctly configured for your deployment environment.
• If you use JWT authentication for APIs, configure require_auth, jwt_secret_key, jwt_expire_time, and jwt_issuer appropriately; MCP tools exposure may depend on API authentication settings.
• When compiling with extension components, follow the recommended go build tags to avoid missing dependencies.
• The MCP configuration is defined in the config.conf file under the [mcp] section; adjust enable, load_components_as_tool, load_chains_as_tool, and load_apis_as_tool as needed.
• If you rename or relocate the UI/editor assets, ensure the server startup path remains correct and compatible with the UI integration instructions in the README.