cua-mcp-server 0.1.7__py3-none-any.whl → 0.1.9__py3-none-any.whl

{cua_mcp_server-0.1.7.dist-info → cua_mcp_server-0.1.9.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: cua-mcp-server
-Version: 0.1.7
+Version: 0.1.9
 Summary: MCP Server for Computer-Use Agent (CUA)
 Author-Email: TryCua <gh@trycua.com>
 Requires-Python: >=3.10
@@ -29,6 +29,17 @@ Description-Content-Type: text/markdown
 **cua-mcp-server** is a MCP server for the Computer-Use Agent (CUA), allowing you to run CUA through Claude Desktop or other MCP clients.
 ### Get started with Agent
 
+## Prerequisites
+
+Before installing the MCP server, you'll need to set up the full Computer-Use Agent capabilities as described in [Option 2 of the main README](../../README.md#option-2-full-computer-use-agent-capabilities). This includes:
+
+1. Installing the Lume CLI
+2. Pulling the latest macOS CUA image
+3. Starting the Lume daemon service
+4. Installing the required Python libraries (Optional: only needed if you want to verify the agent is working before installing MCP server)
+
+Make sure these steps are completed and working before proceeding with the MCP server installation.
+
 ## Installation
 
 Install the package from PyPI:
@@ -68,13 +79,51 @@ You can then use the script in your MCP configuration like this:
         "CUA_AGENT_LOOP": "OMNI",
         "CUA_MODEL_PROVIDER": "ANTHROPIC",
         "CUA_MODEL_NAME": "claude-3-7-sonnet-20250219",
-        "ANTHROPIC_API_KEY": "your-api-key"
+        "CUA_PROVIDER_API_KEY": "your-api-key"
+      }
+    }
+  }
+}
+```
+
+## Development Guide
+
+If you want to develop with the cua-mcp-server directly without installation, you can use this configuration:
+
+```json
+{
+  "mcpServers": {
+    "cua-agent": {
+      "command": "/bin/bash",
+      "args": ["~/cua/libs/mcp-server/scripts/start_mcp_server.sh"],
+      "env": {
+        "CUA_AGENT_LOOP": "UITARS",
+        "CUA_MODEL_PROVIDER": "OAICOMPAT",
+        "CUA_MODEL_NAME": "ByteDance-Seed/UI-TARS-1.5-7B",
+        "CUA_PROVIDER_BASE_URL": "https://****************.us-east-1.aws.endpoints.huggingface.cloud/v1",
+        "CUA_PROVIDER_API_KEY": "your-api-key"
       }
     }
   }
 }
 ```
 
+This configuration:
+- Uses the start_mcp_server.sh script which automatically sets up the Python path and runs the server module
+- Works with Claude Desktop, Cursor, or any other MCP client
+- Automatically uses your development code without requiring installation
+
+Just add this to your MCP client's configuration and it will use your local development version of the server.
+
+### Troubleshooting
+
+If you get a `/bin/bash: ~/cua/libs/mcp-server/scripts/start_mcp_server.sh: No such file or directory` error, try changing the path to the script to be absolute instead of relative.
+
+To see the logs:
+```
+tail -n 20 -f ~/Library/Logs/Claude/mcp*.log
+```
+
 ## Claude Desktop Integration
 
 To use with Claude Desktop, add an entry to your Claude Desktop configuration (`claude_desktop_config.json`, typically found in `~/.config/claude-desktop/`):
@@ -104,7 +153,7 @@ The server is configured using environment variables (can be set in the Claude D
 
 | Variable | Description | Default |
 |----------|-------------|---------|
-| `CUA_AGENT_LOOP` | Agent loop to use (OPENAI, ANTHROPIC, OMNI) | OMNI |
+| `CUA_AGENT_LOOP` | Agent loop to use (OPENAI, ANTHROPIC, UITARS, OMNI) | OMNI |
 | `CUA_MODEL_PROVIDER` | Model provider (ANTHROPIC, OPENAI, OLLAMA, OAICOMPAT) | ANTHROPIC |
 | `CUA_MODEL_NAME` | Model name to use | None (provider default) |
 | `CUA_PROVIDER_BASE_URL` | Base URL for provider API | None |

cua_mcp_server-0.1.9.dist-info/RECORD

@@ -0,0 +1,7 @@
+cua_mcp_server-0.1.9.dist-info/METADATA,sha256=Bw2ET7kbetLRmVqVkyWwZc91NoDSXjn9qNic6pS7T7I,6668
+cua_mcp_server-0.1.9.dist-info/WHEEL,sha256=tSfRZzRHthuv7vxpI4aehrdN9scLjk-dCJkPLzkHxGg,90
+cua_mcp_server-0.1.9.dist-info/entry_points.txt,sha256=Y3uEunDRfoc-RUDS3HnD942RCxYKquiyk-2HRSqphoc,74
+mcp_server/__init__.py,sha256=G5Bps3KxzYfH79B1TDVQI9vbzjamC_mdgi7GJMgbVcA,575
+mcp_server/__main__.py,sha256=BE2ManEiNpz56nqc7Z_asNjQ6TPtvyu5AbWbyJFePnM,132
+mcp_server/server.py,sha256=nV0aNGymSUB1BjwVzS1snUH2phfbVrP3Bl_P_Y4HWII,7907
+cua_mcp_server-0.1.9.dist-info/RECORD,,

mcp_server/server.py

@@ -1,9 +1,10 @@
 import asyncio
+import base64
 import logging
 import os
 import sys
 import traceback
-from typing import Any, Dict, List, Optional, Union
+from typing import Any, Dict, List, Optional, Union, Tuple
 
 # Configure logging to output to stderr for debug visibility
 logging.basicConfig(
@@ -17,7 +18,7 @@ logger = logging.getLogger("mcp-server")
 logger.debug("MCP Server module loading...")
 
 try:
-    from mcp.server.fastmcp import Context, FastMCP
+    from mcp.server.fastmcp import Context, FastMCP, Image
 
     logger.debug("Successfully imported FastMCP")
 except ImportError as e:
@@ -49,16 +50,37 @@ def serve() -> FastMCP:
     server = FastMCP("cua-agent")
 
     @server.tool()
-    async def run_cua_task(ctx: Context, task: str) -> str:
+    async def screenshot_cua(ctx: Context) -> Image:
         """
-        Run a Computer-Use Agent (CUA) task and return the results.
+        Take a screenshot of the current MacOS VM screen and return the image. Use this before running a CUA task to get a snapshot of the current state.
+
+        Args:
+            ctx: The MCP context
+
+        Returns:
+            An image resource containing the screenshot
+        """
+        global global_computer
+        if global_computer is None:
+            global_computer = Computer(verbosity=logging.INFO)
+            await global_computer.run()
+        screenshot = await global_computer.interface.screenshot()
+        return Image(
+            format="png",
+            data=screenshot
+        )
+
+    @server.tool()
+    async def run_cua_task(ctx: Context, task: str) -> Tuple[str, Image]:
+        """
+        Run a Computer-Use Agent (CUA) task in a MacOS VM and return the results.
 
         Args:
             ctx: The MCP context
             task: The instruction or task for the agent to perform
 
         Returns:
-            A string containing the agent's response
+            A tuple containing the agent's response and the final screenshot
         """
         global global_computer
 
@@ -72,12 +94,7 @@ def serve() -> FastMCP:
 
             # Determine which loop to use
             loop_str = os.getenv("CUA_AGENT_LOOP", "OMNI")
-            if loop_str == "OPENAI":
-                loop = AgentLoop.OPENAI
-            elif loop_str == "ANTHROPIC":
-                loop = AgentLoop.ANTHROPIC
-            else:
-                loop = AgentLoop.OMNI
+            loop = getattr(AgentLoop, loop_str)
 
             # Determine provider
             provider_str = os.getenv("CUA_MODEL_PROVIDER", "ANTHROPIC")
@@ -89,6 +106,9 @@ def serve() -> FastMCP:
             # Get base URL for provider (if needed)
             provider_base_url = os.getenv("CUA_PROVIDER_BASE_URL", None)
 
+            # Get api key for provider (if needed)
+            api_key = os.getenv("CUA_PROVIDER_API_KEY", None)
+
             # Create agent with the specified configuration
             agent = ComputerAgent(
                 computer=global_computer,
@@ -98,6 +118,7 @@ def serve() -> FastMCP:
                     name=model_name,
                     provider_base_url=provider_base_url,
                 ),
+                api_key=api_key,
                 save_trajectory=False,
                 only_n_most_recent_images=int(os.getenv("CUA_MAX_IMAGES", "3")),
                 verbosity=logging.INFO,
@@ -107,33 +128,34 @@ def serve() -> FastMCP:
             full_result = ""
             async for result in agent.run(task):
                 logger.info(f"Agent step complete: {result.get('id', 'unknown')}")
+                ctx.info(f"Agent step complete: {result.get('id', 'unknown')}")
 
                 # Add response ID to output
                 full_result += f"\n[Response ID: {result.get('id', 'unknown')}]\n"
-
-                # Extract and concatenate text responses
-                if "text" in result:
-                    # Handle both string and dict responses
-                    text_response = result.get("text", "")
-                    if isinstance(text_response, str):
-                        full_result += f"Response: {text_response}\n"
-                    else:
-                        # If it's a dict or other structure, convert to string representation
-                        full_result += f"Response: {str(text_response)}\n"
-
-                # Log detailed information
-                if "tools" in result:
-                    tools_info = result.get("tools")
-                    logger.debug(f"Tools used: {tools_info}")
-                    full_result += f"\nTools used: {tools_info}\n"
+                
+                if "content" in result:
+                    full_result += f"Response: {result.get('content', '')}\n"
 
                 # Process output if available
                 outputs = result.get("output", [])
                 for output in outputs:
                     output_type = output.get("type")
-                    if output_type == "reasoning":
+                    if output_type == "message":
+                        logger.debug(f"Message: {output}")
+                        content = output.get("content", [])
+                        for content_part in content:
+                            if content_part.get("text"):
+                                full_result += f"\nMessage: {content_part.get('text', '')}\n"
+                    elif output_type == "reasoning":
                         logger.debug(f"Reasoning: {output}")
-                        full_result += f"\nReasoning: {output.get('content', '')}\n"
+                        
+                        summary_content = output.get("summary", [])
+                        if summary_content:
+                            for summary_part in summary_content:
+                                if summary_part.get("text"):
+                                    full_result += f"\nReasoning: {summary_part.get('text', '')}\n"
+                        else:
+                            full_result += f"\nReasoning: {output.get('text', output.get('content', ''))}\n"
                     elif output_type == "computer_call":
                         logger.debug(f"Computer call: {output}")
                         action = output.get("action", "")
@@ -144,17 +166,25 @@ def serve() -> FastMCP:
                 full_result += "\n" + "-" * 40 + "\n"
 
             logger.info(f"CUA task completed successfully")
-            return full_result or "Task completed with no text output."
+            ctx.info(f"CUA task completed successfully")
+            return (
+                full_result or "Task completed with no text output.",
+                Image(
+                    format="png",
+                    data=await global_computer.interface.screenshot()
+                )
+            )
 
         except Exception as e:
             error_msg = f"Error running CUA task: {str(e)}\n{traceback.format_exc()}"
             logger.error(error_msg)
+            ctx.error(error_msg)
             return f"Error during task execution: {str(e)}"
 
     @server.tool()
-    async def run_multi_cua_tasks(ctx: Context, tasks: List[str]) -> str:
+    async def run_multi_cua_tasks(ctx: Context, tasks: List[str]) -> List:
         """
-        Run multiple CUA tasks in sequence and return the combined results.
+        Run multiple CUA tasks in a MacOS VM in sequence and return the combined results.
 
         Args:
             ctx: The MCP context
@@ -164,13 +194,15 @@ def serve() -> FastMCP:
             Combined results from all tasks
         """
         results = []
-
         for i, task in enumerate(tasks):
             logger.info(f"Running task {i+1}/{len(tasks)}: {task}")
-            result = await run_cua_task(ctx, task)
-            results.append(f"Task {i+1}: {task}\nResult: {result}\n")
-
-        return "\n".join(results)
+            ctx.info(f"Running task {i+1}/{len(tasks)}: {task}")
+            
+            ctx.report_progress(i / len(tasks))
+            results.extend(await run_cua_task(ctx, task))
+            ctx.report_progress((i + 1) / len(tasks))
+            
+        return results
 
     return server
 

cua_mcp_server-0.1.7.dist-info/RECORD

@@ -1,7 +0,0 @@
-cua_mcp_server-0.1.7.dist-info/METADATA,sha256=tYO69KlAhGJJBgR3hs3qIggKUnqF-fhebU1Aghgnh-Q,4811
-cua_mcp_server-0.1.7.dist-info/WHEEL,sha256=tSfRZzRHthuv7vxpI4aehrdN9scLjk-dCJkPLzkHxGg,90
-cua_mcp_server-0.1.7.dist-info/entry_points.txt,sha256=Y3uEunDRfoc-RUDS3HnD942RCxYKquiyk-2HRSqphoc,74
-mcp_server/__init__.py,sha256=G5Bps3KxzYfH79B1TDVQI9vbzjamC_mdgi7GJMgbVcA,575
-mcp_server/__main__.py,sha256=BE2ManEiNpz56nqc7Z_asNjQ6TPtvyu5AbWbyJFePnM,132
-mcp_server/server.py,sha256=RdM0kytzt8uF-vbqPXQ3oay-jtGhum4k_Z0jTDZmfoc,6547
-cua_mcp_server-0.1.7.dist-info/RECORD,,