Skip to content

MCP server setup

The Sympheny MCP server exposes the Sympheny API to AI assistants over the Model Context Protocol, so a model can create scenarios, run solver jobs, and read results without you wiring up HTTP calls yourself. It is a remote server, reachable over streamable HTTP.

Not yet available

The MCP server has not been released. The configuration below is the shape you will use once it ships. The server URL is a placeholder and will not connect today. Authentication settings will be added here when the server is published; the exact scheme is not final yet.

Throughout this page the placeholder endpoint is:

https://mcp.sympheny.com/mcp    # placeholder, not live yet

Claude Code

Add the server from the CLI (the --transport http flag is required for a remote server):

claude mcp add --transport http sympheny https://mcp.sympheny.com/mcp

To share the configuration with a team, use --scope project, which writes a checked-in .mcp.json:

{
  "mcpServers": {
    "sympheny": {
      "type": "http",
      "url": "https://mcp.sympheny.com/mcp"
    }
  }
}

Verify the connection with claude mcp get sympheny or the /mcp command.

Claude Desktop

Open Settings → Connectors → Add custom connector, enter the server URL, and click Add. This is the recommended path for remote servers, with no config file to edit.

ChatGPT

Enable Developer mode, then open Settings → Connectors → Create and enter the server URL. When calling a model through the API instead, add the server as an MCP tool:

{
  "type": "mcp",
  "server_label": "sympheny",
  "server_url": "https://mcp.sympheny.com/mcp"
}

Gemini CLI

Add the server to ~/.gemini/settings.json under mcpServers, using httpUrl for the streamable-HTTP transport:

{
  "mcpServers": {
    "sympheny": {
      "httpUrl": "https://mcp.sympheny.com/mcp"
    }
  }
}

Not using an MCP client yet? Drive Sympheny with the REST API or the Python SDK today.