SYSTEM R AI

Configuration

Configure System R MCP server with Claude Desktop, Cursor, and other MCP clients.

The simplest setup. Connect to the hosted MCP server via SSE transport. No local installation required.

Server configuration

{
  "mcpServers": {
    "systemr": {
      "transport": {
        "type": "sse",
        "url": "https://agents.systemr.ai/mcp/sse"
      },
      "metadata": {
        "api_key": "sr_agent_YOUR_KEY_HERE"
      }
    }
  }
}

Claude Desktop

macOS

Edit ~/Library/Application Support/Claude/claude_desktop_config.json:

{
  "mcpServers": {
    "systemr": {
      "transport": {
        "type": "sse",
        "url": "https://agents.systemr.ai/mcp/sse"
      },
      "metadata": {
        "api_key": "sr_agent_YOUR_KEY_HERE"
      }
    }
  }
}

Windows

Edit %APPDATA%\Claude\claude_desktop_config.json with the same configuration.

After saving, restart Claude Desktop. System R tools will appear in the tool list.

Cursor

Add to your Cursor MCP configuration (.cursor/mcp.json in your project root or global settings):

{
  "mcpServers": {
    "systemr": {
      "transport": {
        "type": "sse",
        "url": "https://agents.systemr.ai/mcp/sse"
      },
      "metadata": {
        "api_key": "sr_agent_YOUR_KEY_HERE"
      }
    }
  }
}

Local stdio transport

For development or air-gapped environments, you can run the MCP server locally.

Requirements

pip install mcp systemr

Run the server

python -m src.mcp_server.server

The server starts in stdio mode. Configure your MCP client to use stdio transport:

{
  "mcpServers": {
    "systemr-local": {
      "command": "python",
      "args": ["-m", "src.mcp_server.server"],
      "env": {
        "SYSTEMR_API_KEY": "sr_agent_YOUR_KEY_HERE"
      }
    }
  }
}

Verifying the connection

Once configured, ask your AI assistant to list available tools:

"What System R tools do you have access to?"

The assistant should respond with a list of 55 tools across core, analysis, intelligence, planning, data, system, compound, journal, and memory/ML categories.

Calling a tool via MCP

Once connected, the AI assistant can call tools naturally:

"Calculate position size for AAPL. I want to buy at $185.50 with a stop at $180. My account has $100,000."

The assistant will call calculate_position_size via MCP and return the result.

Authentication

The API key is passed in the MCP session metadata, not as a URL parameter or header. This keeps the key out of URLs and logs.

If the API key is invalid or the agent is inactive, tool calls will return an authentication error through the MCP error response mechanism.

Troubleshooting

IssueSolution
Tools not appearingRestart the MCP client after saving config.
Authentication errorsVerify the API key starts with sr_agent_ and is valid.
Connection timeoutCheck network connectivity to agents.systemr.ai.
"Insufficient balance"Deposit compute credits via the billing endpoints.
System R AI
Python SDKpip install systemr
MCP Serveragents.systemr.ai/mcp/sse
OpenAPI Specagents.systemr.ai/openapi.json
Machine Docsagents.systemr.ai/llms.txt
GitHubSystem-R-AI
X@Systemrai
YouTube@systemr_ai
Emailhello@systemr.ai
Phone628 333 6693
Address7901 4TH ST N, STE 28529, ST PETERSBURG, FL 33702
TermsTerms of Service
PrivacyPrivacy Policy
SecuritySecurity Policy
© 2026 System R AI. Software platform. Not financial advice.

On this page