Improve OmniParser deployment and configuration options

- Add auto-deploy functionality with user confirmation
- Add skip-confirmation flag to deploy without prompting
- Add TODO for simplified AWS configuration in the future
- Update documentation with new options and deployment scenarios
- Expand README with detailed OmniParser configuration instructions

🤖 Generated with [Claude Code](https://claude.ai/code)

Co-Authored-By: Claude <[email protected]>

Author: abrichr

omnimcp/README.md

@@ -72,9 +72,15 @@ omnimcp debug
 # Connect to a remote OmniParser server
 omnimcp cli --server-url=https://your-omniparser-server.example.com
 
+# Deploy OmniParser automatically without confirming
+omnimcp cli --auto-deploy-parser --skip-confirmation
+
 # Allow running even if OmniParser isn't available (limited functionality)
 omnimcp cli --allow-no-parser
 
+# Disable automatic OmniParser deployment attempt
+omnimcp cli --auto-deploy-parser=False
+
 # With additional options
 omnimcp cli --use-normalized-coordinates
 omnimcp debug --debug-dir=/path/to/debug/folder
@@ -89,11 +95,27 @@ OmniMCP requires access to an OmniParser server for analyzing screenshots:
    omnimcp cli --server-url=https://your-omniparser-server.example.com
    ```
 
-2. **Use the Default Local Server**
+2. **Auto-Deploy OmniParser** (Convenient but requires AWS credentials)
+   - By default, OmniMCP will offer to deploy OmniParser if not available
+   - You can control this behavior with these flags:
+     ```bash
+     # Deploy without asking for confirmation
+     omnimcp cli --auto-deploy-parser --skip-confirmation
+     
+     # Disable auto-deployment completely
+     omnimcp cli --auto-deploy-parser=False
+     ```
+
+3. **Use the Default Local Server**
    - OmniMCP will try to connect to `http://localhost:8000` by default
    - This requires running an OmniParser server locally
 
-By default, OmniMCP will fail if it can't connect to an OmniParser server. Use the `--allow-no-parser` flag to run with limited functionality when no parser is available.
+4. **Run Without OmniParser** (Limited functionality)
+   - Use the `--allow-no-parser` flag to run even without OmniParser
+   - Claude will only see raw screenshots without UI element detection
+   ```bash
+   omnimcp cli --allow-no-parser
+   ```
 
 ### Future Direction: Anthropic ComputerUse Integration
 

openadapt/omnimcp.py

@@ -303,7 +303,9 @@ def __init__(
         server_url: Optional[str] = None,
         claude_api_key: Optional[str] = None,
         use_normalized_coordinates: bool = False,
-        allow_no_parser: bool = False
+        allow_no_parser: bool = False,
+        auto_deploy_parser: bool = True,
+        skip_confirmation: bool = False
     ):
         """Initialize OmniMCP.
         
@@ -312,6 +314,8 @@ def __init__(
             claude_api_key: API key for Claude (overrides config)
             use_normalized_coordinates: If True, use normalized (0-1) coordinates
             allow_no_parser: If True, continue even if OmniParser is not available
+            auto_deploy_parser: If True, attempt to deploy OmniParser if not available
+            skip_confirmation: If True, skip user confirmation for deployment
         """
         self.omniparser = OmniParserProvider(server_url)
         self.visual_state = VisualState()
@@ -330,13 +334,39 @@ def __init__(
 
         # Ensure OmniParser is running
         if not self.omniparser.is_available():
-            logger.info("OmniParser not available, attempting to deploy...")
-            self.omniparser.deploy()
+            # Inform user about missing OmniParser
+            if auto_deploy_parser:
+                # Get user confirmation if needed
+                deploy_confirmed = skip_confirmation
+                # TODO: Implement a simplified AWS configuration process
+                # Create an OpenAdapt.AI API key generation system that eliminates the need
+                # for users to manually configure AWS_SECRET_ACCESS_KEY and AWS_ACCESS_ID
+                if not skip_confirmation:
+                    user_input = input(
+                        "\nOmniParser is not available. Would you like to deploy it now? [y/N]: "
+                    ).lower()
+                    deploy_confirmed = user_input in ["y", "yes"]
+                
+                # Attempt to deploy OmniParser if confirmed
+                if deploy_confirmed:
+                    logger.info("Deploying OmniParser service...")
+                    deploy_success = self.omniparser.deploy()
+                    if deploy_success:
+                        logger.info("OmniParser deployed successfully.")
+                    else:
+                        logger.error("Failed to deploy OmniParser.")
+                elif not allow_no_parser:
+                    # User declined deployment and allow_no_parser isn't set
+                    raise RuntimeError(
+                        "OmniParser deployment was declined. Please ensure it's running, "
+                        "use --auto-deploy-parser, or use --allow-no-parser flag."
+                    )
 
-            # Check again after deployment attempt
+            # Final check after deployment attempt
             if not self.omniparser.is_available() and not allow_no_parser:
                 raise RuntimeError(
-                    "OmniParser server is not available. Please ensure it's running or use --allow-no-parser flag."
+                    "OmniParser server is not available. Please ensure it's running, "
+                    "use --auto-deploy-parser, or use --allow-no-parser flag."
                 )
 
     def update_visual_state(self) -> VisualState:

openadapt/run_omnimcp.py

@@ -68,7 +68,9 @@ def cli(
         claude_api_key=None,
         use_normalized_coordinates=False,
         debug_dir=None,
-        allow_no_parser=False
+        allow_no_parser=False,
+        auto_deploy_parser=True,
+        skip_confirmation=False
     ):
         """Run OmniMCP in CLI mode.
         
@@ -87,13 +89,17 @@ def cli(
             use_normalized_coordinates: Use normalized (0-1) coordinates instead of pixels
             debug_dir: Directory to save debug visualizations
             allow_no_parser: If True, continue even if OmniParser is not available
+            auto_deploy_parser: If True, attempt to deploy OmniParser if not available (default: True)
+            skip_confirmation: If True, skip user confirmation for OmniParser deployment
         """
         # Create OmniMCP instance
         omnimcp = OmniMCP(
             server_url=server_url,
             claude_api_key=claude_api_key,  # Will use config.ANTHROPIC_API_KEY if None
             use_normalized_coordinates=use_normalized_coordinates,
-            allow_no_parser=allow_no_parser
+            allow_no_parser=allow_no_parser,
+            auto_deploy_parser=auto_deploy_parser,
+            skip_confirmation=skip_confirmation
         )
 
         # Handle debug directory if specified
@@ -119,7 +125,9 @@ def server(
         claude_api_key=None,
         use_normalized_coordinates=False,
         debug_dir=None,
-        allow_no_parser=False
+        allow_no_parser=False,
+        auto_deploy_parser=True,
+        skip_confirmation=False
     ):
         """Run OmniMCP as an MCP server.
         
@@ -140,13 +148,17 @@ def server(
             use_normalized_coordinates: Use normalized (0-1) coordinates instead of pixels
             debug_dir: Directory to save debug visualizations
             allow_no_parser: If True, continue even if OmniParser is not available
+            auto_deploy_parser: If True, attempt to deploy OmniParser if not available (default: True)
+            skip_confirmation: If True, skip user confirmation for OmniParser deployment
         """
         # Create OmniMCP instance
         omnimcp = OmniMCP(
             server_url=server_url,
             claude_api_key=claude_api_key,  # Will use config.ANTHROPIC_API_KEY if None
             use_normalized_coordinates=use_normalized_coordinates,
-            allow_no_parser=allow_no_parser
+            allow_no_parser=allow_no_parser,
+            auto_deploy_parser=auto_deploy_parser,
+            skip_confirmation=skip_confirmation
         )
 
         # Handle debug directory if specified
@@ -172,7 +184,9 @@ def debug(
         claude_api_key=None,
         use_normalized_coordinates=False,
         debug_dir=None,
-        allow_no_parser=False
+        allow_no_parser=False,
+        auto_deploy_parser=True,
+        skip_confirmation=False
     ):
         """Run OmniMCP in debug mode.
         
@@ -190,13 +204,17 @@ def debug(
             use_normalized_coordinates: Use normalized (0-1) coordinates instead of pixels
             debug_dir: Directory to save debug visualizations
             allow_no_parser: If True, continue even if OmniParser is not available
+            auto_deploy_parser: If True, attempt to deploy OmniParser if not available (default: True)
+            skip_confirmation: If True, skip user confirmation for OmniParser deployment
         """
         # Create OmniMCP instance
         omnimcp = OmniMCP(
             server_url=server_url,
             claude_api_key=claude_api_key,  # Will use config.ANTHROPIC_API_KEY if None
             use_normalized_coordinates=use_normalized_coordinates,
-            allow_no_parser=allow_no_parser
+            allow_no_parser=allow_no_parser,
+            auto_deploy_parser=auto_deploy_parser,
+            skip_confirmation=skip_confirmation
         )
 
         # Create debug directory if not specified