Unstructured API MCP Server

Instructions

Args:
    workflow_id: ID of the workflow to run

Returns:
    String containing the response from the workflow execution

Implementation Reference

• uns_mcp/server.py:391-410 (handler)

  The main handler and registration for the 'run_workflow' tool. This async function is decorated with @mcp.tool() which registers it with the MCP server. It executes the tool logic by calling the UnstructuredClient's run_workflow_async method using the provided workflow_id, handling exceptions and returning the response or error message.

  @mcp.tool()
  async def run_workflow(ctx: Context, workflow_id: str) -> str:
      """Run a specific workflow.
  
      Args:
          workflow_id: ID of the workflow to run
  
      Returns:
          String containing the response from the workflow execution
      """
      client = ctx.request_context.lifespan_context.client
  
      try:
          response = await client.workflows.run_workflow_async(
              request=RunWorkflowRequest(workflow_id=workflow_id),
          )
          return f"Workflow execution initiated: {response.raw_response}"
      except Exception as e:
          return f"Error running workflow: {str(e)}"