MCP Server Setup

Jetty's MCP (Model Context Protocol) server lets any MCP-compatible agent or editor interact with your Jetty workflows. MCP is an open standard that gives AI agents structured access to external tools — in this case, Jetty's full API for creating, running, and monitoring workflows.

Prerequisites

• A Jetty account and API token (see Setup)
• Node.js 18+

Setup by Client

Claude Code

The recommended path for Claude Code is the plugin. If you prefer MCP-only (no slash commands):

Add to your project's .mcp.json:

{
  "mcpServers": {
    "jetty": {
      "command": "npx",
      "args": ["-y", "jetty-mcp-server"],
      "env": {
        "JETTY_API_TOKEN": "mlc_your_token_here"
      }
    }
  }
}

Cursor

Add to .cursor/mcp.json:

{
  "mcpServers": {
    "jetty": {
      "command": "npx",
      "args": ["-y", "jetty-mcp-server"],
      "env": {
        "JETTY_API_TOKEN": "mlc_your_token_here"
      }
    }
  }
}

VS Code Copilot

Add to .vscode/mcp.json:

{
  "servers": {
    "jetty": {
      "command": "npx",
      "args": ["-y", "jetty-mcp-server"],
      "env": {
        "JETTY_API_TOKEN": "mlc_your_token_here"
      }
    }
  }
}

Windsurf

Add to ~/.codeium/windsurf/mcp_config.json:

{
  "mcpServers": {
    "jetty": {
      "command": "npx",
      "args": ["-y", "jetty-mcp-server"],
      "env": {
        "JETTY_API_TOKEN": "mlc_your_token_here"
      }
    }
  }
}

Zed

Add to ~/.config/zed/settings.json under "context_servers":

{
  "context_servers": {
    "jetty": {
      "command": {
        "path": "npx",
        "args": ["-y", "jetty-mcp-server"],
        "env": {
          "JETTY_API_TOKEN": "mlc_your_token_here"
        }
      }
    }
  }
}

Gemini CLI

Install the extension:

gemini extensions install gemini-extension.json

(The gemini-extension.json is included in the agent-skill repo.)

Any MCP Client

Run the server directly:

JETTY_API_TOKEN=mlc_your_token_here npx -y jetty-mcp-server

Environment Variables

Variable	Required	Default	Description
JETTY_API_TOKEN	Yes	—	Your Jetty API token (mlc_...)
JETTY_DOCK_URL	No	https://dock.jetty.io	Dock API base URL
JETTY_FLOWS_URL	No	https://flows-api.jetty.io	Flows API base URL

Available Tools

The MCP server exposes 14 tools:

Collections

Tool	Description
list-collections	List all collections accessible with your token
get-collection	Get collection details including environment variable keys

Tasks

Tool	Description
list-tasks	List tasks in a collection
get-task	Get task details including the workflow JSON definition
create-task	Create a new task with a workflow definition
update-task	Update a task's workflow definition or description

Workflow Execution

Tool	Description
run-workflow	Run a workflow asynchronously — returns a workflow_id immediately
run-workflow-sync	Run a workflow synchronously — blocks until completion (30–60s timeout)

Trajectories

Tool	Description
list-trajectories	List recent workflow runs with pagination
get-trajectory	Get full details of a specific run (inputs, outputs, step results)

Analytics & Labels

Tool	Description
get-stats	Execution statistics for a task
add-label	Add a key-value label to a trajectory (e.g., quality=high)

Step Templates

Tool	Description
list-step-templates	List all available workflow step templates
get-step-template	Get the full schema and documentation for a specific template

Verifying Your Setup

After configuring the MCP server, ask your agent:

"List my Jetty collections"

If the token is valid, you'll see your collections. If you get a 401 error, double-check the token value.

Next Steps

• Agent Recipes — Practical workflows to try with your agent
• Common Gotchas — Known quirks that trip up agents
• Claude Code Plugin — If you're using Claude Code, the plugin adds /jetty commands on top of MCP