MCP Server Setup

The IntelliRag MCP server connects your AI coding assistant to your indexed codebase. It runs as a local process, communicates with the assistant over stdio (JSON-RPC), and calls the IntelliRag API to retrieve code intelligence.

Prerequisites

• An IntelliRag API key (create one in the dashboard)
• The indexer has run at least once against your repository
• Go 1.23+ (for go install) or a prebuilt binary

Install the MCP server

1. Install the binary

   go install github.com/intellirag/intellirag.io/mcp/cmd/server@latest

   This installs the irag-mcp binary to your $GOPATH/bin directory. Alternatively, download a prebuilt binary from the releases page.

2. Configure your AI assistant

   Claude Code

   Add IntelliRag as an MCP server using the CLI:

   claude mcp add intellirag -- env RAG_API_URL=https://mcp.intellirag.io RAG_API_KEY=your-api-key irag-mcp

   Or add it manually to .claude/settings.json:

   {
     "mcpServers": {
       "intellirag": {
         "command": "irag-mcp",
         "env": {
           "RAG_API_URL": "https://mcp.intellirag.io",
           "RAG_API_KEY": "your-api-key"
         }
       }
     }
   }

   Cursor

   Add to .cursor/mcp.json in your project root:

   {
     "mcpServers": {
       "intellirag": {
         "command": "irag-mcp",
         "env": {
           "RAG_API_URL": "https://mcp.intellirag.io",
           "RAG_API_KEY": "your-api-key"
         }
       }
     }
   }

   Windsurf

   Add to your Windsurf MCP configuration file (~/.codeium/windsurf/mcp_config.json):

   {
     "mcpServers": {
       "intellirag": {
         "command": "irag-mcp",
         "env": {
           "RAG_API_URL": "https://mcp.intellirag.io",
           "RAG_API_KEY": "your-api-key"
         }
       }
     }
   }

   VS Code

   Add to .vscode/mcp.json in your project root:

   {
     "servers": {
       "intellirag": {
         "command": "irag-mcp",
         "env": {
           "RAG_API_URL": "https://mcp.intellirag.io",
           "RAG_API_KEY": "your-api-key"
         }
       }
     }
   }

3. Verify the connection

   Ask your AI assistant:

   “What MCP tools do you have from IntelliRag?”

   The assistant should list all 20 IntelliRag tools across five categories: Understanding, Navigation, Impact Analysis, Integration, and Generation.

Environment variables

Variable	Required	Description
RAG_API_URL	Yes	The IntelliRag API server URL (e.g. https://mcp.intellirag.io)
RAG_API_KEY	Yes	Your IntelliRag API key for authentication

Note

All MCP server logging goes to stderr. The stdout channel is reserved for JSON-RPC communication with the AI assistant. If you need to debug the MCP server, check stderr output.

Troubleshooting

“No tools found” or assistant does not list IntelliRag tools

• Verify irag-mcp is in your PATH by running irag-mcp in a terminal
• Check that RAG_API_URL and RAG_API_KEY are set correctly
• Restart your AI assistant after adding the MCP configuration

“Unauthorized” errors in tool responses

• Verify your API key is valid in the dashboard
• Check that the API key has access to the workspace and repository you are querying

Empty results from tools

• Confirm the indexer has run against your repository at least once
• Check the repository status in the dashboard to verify indexing completed successfully

Next steps

• Understanding Tools - semantic search, symbol lookup, module summaries
• Navigation Tools - call graphs, dependency graphs, data flow
• Impact Analysis Tools - change impact, dead code, complexity