Cloudera AI Agent Studio MCP Server

Agent Studio MCP Server is a lightweight Model Context Protocol (MCP) bridge that exposes your Agent Studio instance as a set of callable tools. It lets Claude (or any other MCP-aware client) list, inspect, and build Agent Studio workflows on the fly.

✨ What it can do

Tool	Description	Typical Usage
list_current_workflows()	Returns every workflow’s ID and name	Populate a UI drop-down of existing projects
get_workflow_information(id)	Fetch full JSON metadata for one workflow	Inspect tasks, agents, and process settings
create_workflow(name, description)	Spin up a brand-new blank workflow	Automate project scaffolding
make_workflow_conversational(workflow_id)	Adds a Conversational Task and flips is_conversational to true	Turn a static workflow into a chat-first experience
add_manager_agent_to_workflow(workflow_id, …)	Creates a manager agent and wires it in as the workflow’s supervisor	Hierarchical / delegated workflows
add_agent_to_workflow(workflow_id, …)	Adds any number of additional agents to a workflow	Expand the crew with domain specialists

🖥 Integrating with Claude Desktop

Add the following block to your claude_desktop_config.json:

{
  "mcpServers": {
    "agent-studio": {
      "command": "uv",
      "args": [
        "--directory",
        "/path/to/cloudera-ai-agent-studio-mcp",
        "run",
        "serve.py"
      ],
      "env": {
        "CDSW_APIV2_KEY": "YOUR-TOKEN-HERE",
        "AGENT_STUDIO_DOMAIN": "https://your-studio.CDSW_DOMAIN.cldr.work"
      }
    }
  }
}