cua-agent 0.1.2__py3-none-any.whl → 0.1.4__py3-none-any.whl

agent/__init__.py

@@ -1,7 +1,53 @@
 """CUA (Computer Use) Agent for AI-driven computer interaction."""
 
+import sys
+import logging
+
 __version__ = "0.1.0"
 
+# Initialize logging
+logger = logging.getLogger("cua.agent")
+
+# Initialize telemetry when the package is imported
+try:
+    # Import from core telemetry for basic functions
+    from core.telemetry import (
+        is_telemetry_enabled,
+        flush,
+        record_event,
+    )
+
+    # Import set_dimension from our own telemetry module
+    from .core.telemetry import set_dimension
+
+    # Check if telemetry is enabled
+    if is_telemetry_enabled():
+        logger.info("Telemetry is enabled")
+
+        # Record package initialization
+        record_event(
+            "module_init",
+            {
+                "module": "agent",
+                "version": __version__,
+                "python_version": sys.version,
+            },
+        )
+
+        # Set the package version as a dimension
+        set_dimension("agent_version", __version__)
+
+        # Flush events to ensure they're sent
+        flush()
+    else:
+        logger.info("Telemetry is disabled")
+except ImportError as e:
+    # Telemetry not available
+    logger.warning(f"Telemetry not available: {e}")
+except Exception as e:
+    # Other issues with telemetry
+    logger.warning(f"Error initializing telemetry: {e}")
+
 from .core.factory import AgentFactory
 from .core.agent import ComputerAgent
 from .providers.omni.types import LLMProvider, LLM

agent/core/agent.py

@@ -3,13 +3,17 @@
 import os
 import logging
 import asyncio
+import time
+import uuid
 from typing import Any, AsyncGenerator, Dict, List, Optional, TYPE_CHECKING, Union, cast
 from datetime import datetime
+from enum import Enum
 
 from computer import Computer
 
 from ..types.base import Provider, AgentLoop
 from .base_agent import BaseComputerAgent
+from ..core.telemetry import record_agent_initialization
 
 # Only import types for type checking to avoid circular imports
 if TYPE_CHECKING:
@@ -26,13 +30,11 @@ logger = logging.getLogger(__name__)
 DEFAULT_MODELS = {
     LLMProvider.OPENAI: "gpt-4o",
     LLMProvider.ANTHROPIC: "claude-3-7-sonnet-20250219",
-    LLMProvider.GROQ: "llama3-70b-8192",
 }
 
 # Map providers to their environment variable names
 ENV_VARS = {
     LLMProvider.OPENAI: "OPENAI_API_KEY",
-    LLMProvider.GROQ: "GROQ_API_KEY",
     LLMProvider.ANTHROPIC: "ANTHROPIC_API_KEY",
 }
 
@@ -55,79 +57,61 @@ class ComputerAgent(BaseComputerAgent):
         only_n_most_recent_images: Optional[int] = None,
         max_retries: int = 3,
         verbosity: int = logging.INFO,
+        telemetry_enabled: bool = True,
         **kwargs,
     ):
-        """Initialize the computer agent.
+        """Initialize a ComputerAgent instance.
 
         Args:
-            computer: Computer instance to control
-            loop: The type of loop to use (Anthropic or Omni)
-            model: LLM configuration. Can be:
-                  - LLM object with provider and name
-                  - Dict with 'provider' and 'name' keys
-                  - String with model name (defaults to OpenAI provider)
-                  - None (defaults based on loop)
-            api_key: Optional API key (will use environment variable if not provided)
-            save_trajectory: Whether to save screenshots and logs
-            trajectory_dir: Directory to save trajectories (defaults to "trajectories")
-            only_n_most_recent_images: Limit history to N most recent images
-            max_retries: Maximum number of retry attempts for failed operations
-            verbosity: Logging level (standard Python logging levels: logging.DEBUG, logging.INFO, etc.)
-            **kwargs: Additional keyword arguments to pass to the loop
+            computer: The Computer instance to control
+            loop: The agent loop to use: ANTHROPIC or OMNI
+            model: The model to use. Can be a string, dict or LLM object.
+                  Defaults to LLM for the loop type.
+            api_key: The API key to use. If None, will use environment variables.
+            save_trajectory: Whether to save the trajectory.
+            trajectory_dir: The directory to save trajectories to.
+            only_n_most_recent_images: Only keep this many most recent images.
+            max_retries: Maximum number of retries for failed requests.
+            verbosity: Logging level (standard Python logging levels).
+            telemetry_enabled: Whether to enable telemetry tracking. Defaults to True.
+            **kwargs: Additional keyword arguments to pass to the loop.
         """
-        # Set up trajectory directories based on save_trajectory
-        base_dir = trajectory_dir if save_trajectory else None
-        # Don't create a redundant screenshots directory - directly use the timestamp folder
-        screenshot_dir = None  # This was previously set to os.path.join(base_dir, "screenshots")
-        log_dir = None
-
-        super().__init__(
-            max_retries=max_retries,
-            computer=computer,
-            screenshot_dir=screenshot_dir,
-            log_dir=log_dir,
-            **kwargs,
-        )
+        super().__init__(computer)
+        self._configure_logging(verbosity)
+        logger.info(f"Initializing ComputerAgent with {loop} loop")
 
+        # Store telemetry preference
+        self.telemetry_enabled = telemetry_enabled
+
+        # Process the model configuration
+        self.model = self._process_model_config(model, loop)
         self.loop_type = loop
+        self.api_key = api_key
+
+        # Store computer
+        self.computer = computer
+
+        # Save trajectory settings
         self.save_trajectory = save_trajectory
         self.trajectory_dir = trajectory_dir
         self.only_n_most_recent_images = only_n_most_recent_images
-        self.verbosity = verbosity
-        self._kwargs = kwargs  # Keep this for loop initialization
 
-        # Configure logging based on verbosity
-        self._configure_logging(verbosity)
+        # Store the max retries setting
+        self.max_retries = max_retries
 
-        # Process model configuration
-        self.model_config = self._process_model_config(model, loop)
+        # Initialize message history
+        self.messages = []
 
-        # Get API key from environment if not provided
-        if api_key is None:
-            env_var = (
-                ENV_VARS.get(self.model_config.provider)
-                if loop == AgentLoop.OMNI
-                else "ANTHROPIC_API_KEY"
-            )
-            if not env_var:
-                raise ValueError(
-                    f"Unsupported provider: {self.model_config.provider}. Please use one of: {list(ENV_VARS.keys())}"
-                )
-
-            api_key = os.environ.get(env_var)
-            if not api_key:
-                raise ValueError(
-                    f"No API key provided and {env_var} environment variable is not set.\n"
-                    f"Please set the {env_var} environment variable or pass the api_key directly:\n"
-                    f"  - Export in terminal: export {env_var}=your_api_key_here\n"
-                    f"  - Add to .env file: {env_var}=your_api_key_here\n"
-                    f"  - Pass directly: api_key='your_api_key_here'"
-                )
-        self.api_key = api_key
+        # Extra kwargs for the loop
+        self.loop_kwargs = kwargs
 
-        # Initialize the appropriate loop based on loop_type
+        # Initialize the actual loop implementation
         self.loop = self._init_loop()
 
+        # Record initialization in telemetry if enabled
+        if telemetry_enabled:
+            record_agent_initialization()
+
     def _process_model_config(
         self, model_input: Optional[Union[LLM, Dict[str, str], str]], loop: AgentLoop
     ) -> LLM:
@@ -200,7 +184,7 @@ class ComputerAgent(BaseComputerAgent):
             from ..providers.anthropic.loop import AnthropicLoop
 
             # Ensure we always have a valid model name
-            model_name = self.model_config.name or DEFAULT_MODELS[LLMProvider.ANTHROPIC]
+            model_name = self.model.name or DEFAULT_MODELS[LLMProvider.ANTHROPIC]
 
             return AnthropicLoop(
                 api_key=self.api_key,
@@ -209,119 +193,60 @@ class ComputerAgent(BaseComputerAgent):
                 save_trajectory=self.save_trajectory,
                 base_dir=self.trajectory_dir,
                 only_n_most_recent_images=self.only_n_most_recent_images,
-                **self._kwargs,
+                **self.loop_kwargs,
             )
 
         # Initialize parser for OmniLoop with appropriate device
-        if "parser" not in self._kwargs:
-            self._kwargs["parser"] = OmniParser()
+        if "parser" not in self.loop_kwargs:
+            self.loop_kwargs["parser"] = OmniParser()
 
         # Ensure we always have a valid model name
-        model_name = self.model_config.name or DEFAULT_MODELS[self.model_config.provider]
+        model_name = self.model.name or DEFAULT_MODELS[self.model.provider]
 
         return OmniLoop(
-            provider=self.model_config.provider,
+            provider=self.model.provider,
             api_key=self.api_key,
             model=model_name,
             computer=self.computer,
             save_trajectory=self.save_trajectory,
             base_dir=self.trajectory_dir,
             only_n_most_recent_images=self.only_n_most_recent_images,
-            **self._kwargs,
+            **self.loop_kwargs,
         )
 
     async def _execute_task(self, task: str) -> AsyncGenerator[Dict[str, Any], None]:
-        """Execute a task using the appropriate loop.
+        """Execute a task using the appropriate agent loop.
 
         Args:
-            task: Task description to execute
+            task: The task to execute
 
-        Yields:
-            Dict containing response content and metadata
+        Returns:
+            AsyncGenerator yielding task outputs
         """
+        logger.info(f"Executing task: {task}")
+
         try:
-            # Format the messages based on loop type
-            if self.loop_type == AgentLoop.ANTHROPIC:
-                # Anthropic format
-                messages = [{"role": "user", "content": [{"type": "text", "text": task}]}]
-            else:
-                # Cua format
-                messages = [{"role": "user", "content": task}]
-
-            # Run the loop
-            try:
-                async for result in self.loop.run(messages):
-                    if result is None:
-                        break
-
-                    # Handle error case
-                    if "error" in result:
-                        yield {
-                            "role": "assistant",
-                            "content": result["error"],
-                            "metadata": {"title": "❌ Error"},
-                        }
-                        continue
-
-                    # Extract content and metadata based on loop type
-                    if self.loop_type == AgentLoop.ANTHROPIC:
-                        # Handle Anthropic format
-                        if "content" in result:
-                            content_text = ""
-                            for content_block in result["content"]:
-                                try:
-                                    # Try to access the text attribute directly
-                                    content_text += content_block.text
-                                except (AttributeError, TypeError):
-                                    # If it's a dictionary instead of an object
-                                    if isinstance(content_block, dict) and "text" in content_block:
-                                        content_text += content_block["text"]
-
-                            yield {
-                                "role": "assistant",
-                                "content": content_text,
-                                "metadata": result.get("parsed_screen", {}),
-                            }
-                        else:
-                            yield {
-                                "role": "assistant",
-                                "content": str(result),
-                                "metadata": {"title": "Screen Analysis"},
-                            }
-                    else:
-                        # Handle Omni format
-                        content = ""
-                        metadata = {"title": "Screen Analysis"}
-
-                        # If result has content (normal case)
-                        if "content" in result:
-                            content = result["content"]
-
-                            # Ensure metadata has a title
-                            if isinstance(content, dict) and "metadata" in content:
-                                metadata = content["metadata"]
-                                if "title" not in metadata:
-                                    metadata["title"] = "Screen Analysis"
-
-                            # For string content, convert to proper format
-                            if isinstance(content, str):
-                                content = content
-                            elif isinstance(content, dict) and "content" in content:
-                                content = content.get("content", "")
-
-                        yield {"role": "assistant", "content": content, "metadata": metadata}
-            except Exception as e:
-                logger.error(f"Error running the loop: {str(e)}")
-                yield {
-                    "role": "assistant",
-                    "content": f"Error running the agent loop: {str(e)}",
-                    "metadata": {"title": "❌ Loop Error"},
-                }
+            # Create a message from the task
+            task_message = {"role": "user", "content": task}
+            messages_with_task = self.messages + [task_message]
 
+            # Use the run method of the loop
+            async for output in self.loop.run(messages_with_task):
+                yield output
+        except Exception as e:
+            logger.error(f"Error executing task: {e}")
+            raise
+        finally:
+            pass
+
+    async def _execute_action(self, action_type: str, **action_params) -> Any:
+        """Execute an action with telemetry tracking."""
+        try:
+            # Execute the action
+            result = await super()._execute_action(action_type, **action_params)
+            return result
         except Exception as e:
-            logger.error(f"Error in _execute_task: {str(e)}")
-            yield {
-                "role": "assistant",
-                "content": f"Error: {str(e)}",
-                "metadata": {"title": "❌ Error"},
-            }
+            logger.exception(f"Error executing action {action_type}: {e}")
+            raise
+        finally:
+            pass

agent/core/base_agent.py

@@ -113,7 +113,7 @@ class BaseComputerAgent(ABC):
             # Take a test screenshot to verify the computer is working
             logger.info("Testing computer with a screenshot...")
             try:
-                test_screenshot = await self.computer.screenshot()
+                test_screenshot = await self.computer.interface.screenshot()
                 # Determine the screenshot size based on its type
                 if isinstance(test_screenshot, bytes):
                     size = len(test_screenshot)

agent/core/experiment.py

@@ -8,6 +8,7 @@ from datetime import datetime
 from typing import Any, Dict, List, Optional
 from PIL import Image
 import json
+import re
 
 logger = logging.getLogger(__name__)
 
@@ -106,9 +107,18 @@ class ExperimentManager:
             # Increment screenshot counter
             self.screenshot_count += 1
 
+            # Sanitize action_type to ensure valid filename
+            # Replace characters that are not safe for filenames
+            sanitized_action = ""
+            if action_type:
+                # Replace invalid filename characters with underscores
+                sanitized_action = re.sub(r'[\\/*?:"<>|]', "_", action_type)
+                # Limit the length to avoid excessively long filenames
+                sanitized_action = sanitized_action[:50]
+
             # Create a descriptive filename
             timestamp = int(datetime.now().timestamp() * 1000)
-            action_suffix = f"_{action_type}" if action_type else ""
+            action_suffix = f"_{sanitized_action}" if sanitized_action else ""
             filename = f"screenshot_{self.screenshot_count:03d}{action_suffix}_{timestamp}.png"
 
             # Save directly to the turn directory

agent/core/loop.py

@@ -166,7 +166,7 @@ class BaseLoop(ABC):
         """
         try:
             # Take screenshot
-            screenshot = await self.computer.screenshot()
+            screenshot = await self.computer.interface.screenshot()
 
             # Initialize with default values
             width, height = 1024, 768

agent/core/telemetry.py

@@ -0,0 +1,130 @@
+"""Agent telemetry for tracking anonymous usage and feature usage."""
+
+import logging
+import os
+import platform
+import sys
+from typing import Dict, Any
+
+# Import the core telemetry module
+TELEMETRY_AVAILABLE = False
+
+try:
+    from core.telemetry import (
+        record_event,
+        increment,
+        get_telemetry_client,
+        flush,
+        is_telemetry_enabled,
+        is_telemetry_globally_disabled,
+    )
+
+    def increment_counter(counter_name: str, value: int = 1) -> None:
+        """Wrapper for increment to maintain backward compatibility."""
+        if is_telemetry_enabled():
+            increment(counter_name, value)
+
+    def set_dimension(name: str, value: Any) -> None:
+        """Set a dimension that will be attached to all events."""
+        logger = logging.getLogger("cua.agent.telemetry")
+        logger.debug(f"Setting dimension {name}={value}")
+
+    TELEMETRY_AVAILABLE = True
+    logger = logging.getLogger("cua.agent.telemetry")
+    logger.info("Successfully imported telemetry")
+except ImportError as e:
+    logger = logging.getLogger("cua.agent.telemetry")
+    logger.warning(f"Could not import telemetry: {e}")
+    TELEMETRY_AVAILABLE = False
+
+
+# Local fallbacks in case core telemetry isn't available
+def _noop(*args: Any, **kwargs: Any) -> None:
+    """No-op function for when telemetry is not available."""
+    pass
+
+
+logger = logging.getLogger("cua.agent.telemetry")
+
+# If telemetry isn't available, use no-op functions
+if not TELEMETRY_AVAILABLE:
+    logger.debug("Telemetry not available, using no-op functions")
+    record_event = _noop  # type: ignore
+    increment_counter = _noop  # type: ignore
+    set_dimension = _noop  # type: ignore
+    get_telemetry_client = lambda: None  # type: ignore
+    flush = _noop  # type: ignore
+    is_telemetry_enabled = lambda: False  # type: ignore
+    is_telemetry_globally_disabled = lambda: True  # type: ignore
+
+# Get system info once to use in telemetry
+SYSTEM_INFO = {
+    "os": platform.system().lower(),
+    "os_version": platform.release(),
+    "python_version": platform.python_version(),
+}
+
+
+def enable_telemetry() -> bool:
+    """Enable telemetry if available.
+
+    Returns:
+        bool: True if telemetry was successfully enabled, False otherwise
+    """
+    global TELEMETRY_AVAILABLE
+
+    # Check if globally disabled using core function
+    if TELEMETRY_AVAILABLE and is_telemetry_globally_disabled():
+        logger.info("Telemetry is globally disabled via environment variable - cannot enable")
+        return False
+
+    # Already enabled
+    if TELEMETRY_AVAILABLE:
+        return True
+
+    # Try to import and enable
+    try:
+        from core.telemetry import (
+            record_event,
+            increment,
+            get_telemetry_client,
+            flush,
+            is_telemetry_globally_disabled,
+        )
+
+        # Check again after import
+        if is_telemetry_globally_disabled():
+            logger.info("Telemetry is globally disabled via environment variable - cannot enable")
+            return False
+
+        TELEMETRY_AVAILABLE = True
+        logger.info("Telemetry successfully enabled")
+        return True
+    except ImportError as e:
+        logger.warning(f"Could not enable telemetry: {e}")
+        return False
+
+
+def is_telemetry_enabled() -> bool:
+    """Check if telemetry is enabled.
+
+    Returns:
+        bool: True if telemetry is enabled, False otherwise
+    """
+    # Use the core function if available, otherwise use our local flag
+    if TELEMETRY_AVAILABLE:
+        from core.telemetry import is_telemetry_enabled as core_is_enabled
+
+        return core_is_enabled()
+    return False
+
+
+def record_agent_initialization() -> None:
+    """Record when an agent instance is initialized."""
+    if TELEMETRY_AVAILABLE and is_telemetry_enabled():
+        record_event("agent_initialized", SYSTEM_INFO)
+
+        # Set dimensions that will be attached to all events
+        set_dimension("os", SYSTEM_INFO["os"])
+        set_dimension("os_version", SYSTEM_INFO["os_version"])
+        set_dimension("python_version", SYSTEM_INFO["python_version"])

agent/providers/omni/loop.py

@@ -731,7 +731,7 @@ class OmniLoop(BaseLoop):
                         action_type = f"hotkey_{content['Value'].replace('+', '_')}"
                     logger.info(f"Preparing hotkey with keys: {keys}")
                     # Get the method but call it with *args instead of **kwargs
-                    method = getattr(self.computer, action)
+                    method = getattr(self.computer.interface, action)
                     await method(*keys)  # Unpack the keys list as positional arguments
                     logger.info(f"Tool execution completed successfully: {action}")
 
@@ -776,7 +776,7 @@ class OmniLoop(BaseLoop):
 
                 # Execute tool and handle result
                 try:
-                    method = getattr(self.computer, action)
+                    method = getattr(self.computer.interface, action)
                     logger.info(f"Found method for action '{action}': {method}")
                     await method(**kwargs)
                     logger.info(f"Tool execution completed successfully: {action}")

agent/providers/omni/parser.py

@@ -79,7 +79,7 @@ class OmniParser:
         try:
             # Get screenshot from computer
             logger.info("Taking screenshot...")
-            screenshot = await computer.screenshot()
+            screenshot = await computer.interface.screenshot()
 
             # Log screenshot info
             logger.info(f"Screenshot type: {type(screenshot)}")

agent/providers/omni/types.py

@@ -10,8 +10,6 @@ class LLMProvider(StrEnum):
 
     ANTHROPIC = "anthropic"
     OPENAI = "openai"
-    GROQ = "groq"
-    QWEN = "qwen"
 
 
 LLMProvider
@@ -39,14 +37,10 @@ Model = LLM
 PROVIDER_TO_DEFAULT_MODEL: Dict[LLMProvider, str] = {
     LLMProvider.ANTHROPIC: "claude-3-7-sonnet-20250219",
     LLMProvider.OPENAI: "gpt-4o",
-    LLMProvider.GROQ: "deepseek-r1-distill-llama-70b",
-    LLMProvider.QWEN: "qwen2.5-vl-72b-instruct",
 }
 
 # Environment variable names for each provider
 PROVIDER_TO_ENV_VAR: Dict[LLMProvider, str] = {
     LLMProvider.ANTHROPIC: "ANTHROPIC_API_KEY",
     LLMProvider.OPENAI: "OPENAI_API_KEY",
-    LLMProvider.GROQ: "GROQ_API_KEY",
-    LLMProvider.QWEN: "QWEN_API_KEY",
 }

agent/telemetry.py

@@ -0,0 +1,21 @@
+"""Telemetry support for Agent class."""
+
+import os
+import platform
+import sys
+import time
+from typing import Any, Dict, Optional
+
+from core.telemetry import (
+    record_event,
+    is_telemetry_enabled,
+    flush,
+    get_telemetry_client,
+    increment,
+)
+
+# System information used for telemetry
+SYSTEM_INFO = {
+    "os": sys.platform,
+    "python_version": platform.python_version(),
+}

cua_agent-0.1.4.dist-info/METADATA

@@ -0,0 +1,120 @@
+Metadata-Version: 2.1
+Name: cua-agent
+Version: 0.1.4
+Summary: CUA (Computer Use) Agent for AI-driven computer interaction
+Author-Email: TryCua <gh@trycua.com>
+Requires-Python: <3.13,>=3.10
+Requires-Dist: httpx<0.29.0,>=0.27.0
+Requires-Dist: aiohttp<4.0.0,>=3.9.3
+Requires-Dist: asyncio
+Requires-Dist: anyio<5.0.0,>=4.4.1
+Requires-Dist: typing-extensions<5.0.0,>=4.12.2
+Requires-Dist: pydantic<3.0.0,>=2.6.4
+Requires-Dist: rich<14.0.0,>=13.7.1
+Requires-Dist: python-dotenv<2.0.0,>=1.0.1
+Requires-Dist: cua-computer<0.2.0,>=0.1.0
+Requires-Dist: cua-core<0.2.0,>=0.1.0
+Requires-Dist: certifi>=2024.2.2
+Provides-Extra: anthropic
+Requires-Dist: anthropic>=0.49.0; extra == "anthropic"
+Requires-Dist: boto3<2.0.0,>=1.35.81; extra == "anthropic"
+Provides-Extra: som
+Requires-Dist: torch>=2.2.1; extra == "som"
+Requires-Dist: torchvision>=0.17.1; extra == "som"
+Requires-Dist: ultralytics>=8.0.0; extra == "som"
+Requires-Dist: transformers>=4.38.2; extra == "som"
+Requires-Dist: cua-som<0.2.0,>=0.1.0; extra == "som"
+Requires-Dist: anthropic<0.47.0,>=0.46.0; extra == "som"
+Requires-Dist: boto3<2.0.0,>=1.35.81; extra == "som"
+Requires-Dist: openai<2.0.0,>=1.14.0; extra == "som"
+Requires-Dist: groq<0.5.0,>=0.4.0; extra == "som"
+Requires-Dist: dashscope<2.0.0,>=1.13.0; extra == "som"
+Requires-Dist: requests<3.0.0,>=2.31.0; extra == "som"
+Provides-Extra: all
+Requires-Dist: torch>=2.2.1; extra == "all"
+Requires-Dist: torchvision>=0.17.1; extra == "all"
+Requires-Dist: ultralytics>=8.0.0; extra == "all"
+Requires-Dist: transformers>=4.38.2; extra == "all"
+Requires-Dist: cua-som<0.2.0,>=0.1.0; extra == "all"
+Requires-Dist: anthropic<0.47.0,>=0.46.0; extra == "all"
+Requires-Dist: boto3<2.0.0,>=1.35.81; extra == "all"
+Requires-Dist: openai<2.0.0,>=1.14.0; extra == "all"
+Requires-Dist: groq<0.5.0,>=0.4.0; extra == "all"
+Requires-Dist: dashscope<2.0.0,>=1.13.0; extra == "all"
+Requires-Dist: requests<3.0.0,>=2.31.0; extra == "all"
+Description-Content-Type: text/markdown
+
+<div align="center">
+<h1>
+  <div class="image-wrapper" style="display: inline-block;">
+    <picture>
+      <source media="(prefers-color-scheme: dark)" alt="logo" height="150" srcset="../../img/logo_white.png" style="display: block; margin: auto;">
+      <source media="(prefers-color-scheme: light)" alt="logo" height="150" srcset="../../img/logo_black.png" style="display: block; margin: auto;">
+      <img alt="Shows my svg">
+    </picture>
+  </div>
+
+  [![Python](https://img.shields.io/badge/Python-333333?logo=python&logoColor=white&labelColor=333333)](#)
+  [![macOS](https://img.shields.io/badge/macOS-000000?logo=apple&logoColor=F0F0F0)](#)
+  [![Discord](https://img.shields.io/badge/Discord-%235865F2.svg?&logo=discord&logoColor=white)](https://discord.com/invite/mVnXXpdE85)
+  [![PyPI](https://img.shields.io/pypi/v/cua-computer?color=333333)](https://pypi.org/project/cua-computer/)
+</h1>
+</div>
+
+**Agent** is a Computer Use (CUA) framework for running multi-app agentic workflows targeting macOS and Linux sandbox, supporting local (Ollama) and cloud model providers (OpenAI, Anthropic, Groq, DeepSeek, Qwen). The framework integrates with Microsoft's OmniParser for enhanced UI understanding and interaction.
+
+### Get started with Agent
+
+```python
+from agent import ComputerAgent, AgentLoop, LLMProvider
+from computer import Computer
+
+computer = Computer(verbosity=logging.INFO)
+
+agent = ComputerAgent(
+    computer=computer,
+    loop=AgentLoop.ANTHROPIC,
+    # loop=AgentLoop.OMNI,
+    model=LLM(provider=LLMProvider.ANTHROPIC, name="claude-3-7-sonnet-20250219"),
+    # model=LLM(provider=LLMProvider.OPENAI, name="gpt-4.5-preview"),
+    save_trajectory=True,
+    trajectory_dir=str(Path("trajectories")),
+    only_n_most_recent_images=3,
+    verbosity=logging.INFO,
+)
+
+tasks = [
+"""
+Please help me with the following task:
+1. Open Safari browser
+2. Go to Wikipedia.org
+3. Search for "Claude AI" 
+4. Summarize the main points you find about Claude AI
+"""
+]
+
+async with agent:
+    for i, task in enumerate(tasks, 1):
+        print(f"\nExecuting task {i}/{len(tasks)}: {task}")
+        async for result in agent.run(task):
+            print(result)
+        print(f"Task {i} completed")
+```
+
+## Install
+
+### cua-agent
+
+```bash
+pip install "cua-agent[all]"
+
+# or install specific loop providers
+pip install "cua-agent[anthropic]"
+pip install "cua-agent[omni]"
+```
+
+## Run
+
+Refer to these notebooks for step-by-step guides on how to use the Computer-Use Agent (CUA):
+
+- [Agent Notebook](../../notebooks/agent_nb.ipynb) - Complete examples and workflows

{cua_agent-0.1.2.dist-info → cua_agent-0.1.4.dist-info}/RECORD

@@ -1,15 +1,16 @@
 agent/README.md,sha256=8EFnLrKejthEcL9bZflQSbvA-KwpiPanBz8TEEwRub8,2153
-agent/__init__.py,sha256=BRIunVPG0T5CdAiNJyElKxUZN8Mngg2_TmtLwaupG4I,355
+agent/__init__.py,sha256=sxUp_I8cI4NVa2DY8g-tsJe5_XJJe5VdeS_VLgA98EM,1569
 agent/core/README.md,sha256=VOXNVbR0ugxf9gCXYmZtUU2kngZhfi29haT_oSxK0Lk,3559
 agent/core/__init__.py,sha256=0htZ-VfsH9ixHB8j_SXu_uv6r3XXsq5TrghFNd-yRNE,709
-agent/core/agent.py,sha256=AQ-S2wVD82RFnD_HmR-zjA7Jj09CUKGp7KreWX1j6Fg,13495
-agent/core/base_agent.py,sha256=MgaMKTwgqNJ1-TgS_mxALoC9COzc7Acg9y7Q8HAFX2c,6266
+agent/core/agent.py,sha256=A07a7mRtKqpX2AHCP1i8KesOqoOETfh23CyTTQth6vI,9327
+agent/core/base_agent.py,sha256=te9rk2tJZpEhDUEB1xSaFqe1zeOjmzMdHF5LaUDP2K0,6276
 agent/core/callbacks.py,sha256=VbGIf5QkHh3Q0KsLM6wv7hRdIA5WExTVYLm64bckyUA,4306
 agent/core/computer_agent.py,sha256=JGLMl_PwImUttmQh2amdLlXHS9CUyZ9MW20J1Xid7dM,2417
-agent/core/experiment.py,sha256=AST1t83eqaGzjoW6KvrhfVIs3ELAR_I70VHq2NsMmNk,7446
+agent/core/experiment.py,sha256=FKmSDyA2YFSrO3q-91ZT29Jm1lm24YCuK59wQ6z-6IM,7930
 agent/core/factory.py,sha256=WraOEHWPXBSN4R3DO7M2ctyadodeA8tzHM3dUjdQ_3A,3441
-agent/core/loop.py,sha256=E-0pz7MaguZQrHs5GP98Oc8C_Iz8ier0vXrD9Ny2HL8,8999
+agent/core/loop.py,sha256=vhdlSy_hIY3-a92uTGdF3oYE5Qcq0U2hyTJNmXunnfc,9009
 agent/core/messages.py,sha256=N8pV8Eh-AJpMuDPRI5OGWUIOU6DRr-pQjK9XU0go9Hk,7637
+agent/core/telemetry.py,sha256=cCnITdDxOSHM0qKV7Fe5sV2gD6B_InRxMVFm-EgKF7M,4083
 agent/core/tools/__init__.py,sha256=xZen-PqUp2dUaMEHJowXCQm33_5Sxhsx9PSoD0rq6tI,489
 agent/core/tools/base.py,sha256=CdzRFNuOjNfzgyTUN4ZoCGkUDR5HI0ECQVpvrUdEij8,2295
 agent/core/tools/bash.py,sha256=jnJKVlHn8np8e0gWd8EO0_qqjMkfQzutSugA_Iol4jE,1585
@@ -43,23 +44,24 @@ agent/providers/omni/clients/openai.py,sha256=E4TAXMUFoYTunJETCWCNx5XAc6xutiN4rB
 agent/providers/omni/clients/utils.py,sha256=Ani9CVVBm_J2Dl51WG6p1GVuoI6cq8scISrG0pmQ37o,688
 agent/providers/omni/experiment.py,sha256=JGAdHi7Nf73I48c9k3TY1Xpr_i6D2VG1wurOzw5cNGk,9888
 agent/providers/omni/image_utils.py,sha256=qIFuNi5cIMVwrqYBXG1T6PxUlbxz7gIngFFP39bZIlU,2782
-agent/providers/omni/loop.py,sha256=mHCs13in3mrLizF1x8OeCXECp4bL9-CYS_XOJOUZqu8,43827
+agent/providers/omni/loop.py,sha256=72o7q92nO7i0EUrVhEPCEHprRKdBYsg5iLTLfLHXAsw,43847
 agent/providers/omni/messages.py,sha256=zdjQCAMH-hOyrQQesHhTiIsQbw43KqVSmVIzS8JOIFA,6134
-agent/providers/omni/parser.py,sha256=Iv-cXWG2qzdYjyZJH5pGUzfv6nOaiHQ2OXdQSe00Ydw,9151
+agent/providers/omni/parser.py,sha256=lTAoSMSf2zpwqR_8W0SXG3cYIFeUiZa5vXdpjqZwEHY,9161
 agent/providers/omni/prompts.py,sha256=Mupjy0bUwBjcAeLXpE1r1jisYPSlhwsp-IXJKEKrEtw,3779
 agent/providers/omni/tool_manager.py,sha256=O6DxyEI-Vg6jt99phh011o4q4me_vNhH2YffIxkO4GM,2585
 agent/providers/omni/tools/__init__.py,sha256=l636hx9Q5z9eaFdPanPwPENUE-w-Xm8kAZhPUq0ZQF4,309
 agent/providers/omni/tools/bash.py,sha256=y_ibfP9iRcbiU_E0faAoa4DCP_BlkMlKOOURdBBIGZE,2030
 agent/providers/omni/tools/computer.py,sha256=xkMmAR0e_kbf0Zs2mggCDyWrQOJZyXOKPFjkutaQb94,9108
 agent/providers/omni/tools/manager.py,sha256=V_tav2yU92PyQnFlxNXG1wvNEaJoEYudtKx5sRjj06Q,2619
-agent/providers/omni/types.py,sha256=G7Zqm-nWMa3K2klj-D3KUVWc2r8NJB7sYZCwwl0m9Ic,1233
+agent/providers/omni/types.py,sha256=rpr7-mH9VK1R-nJ6tVu1gKp427j-hw1DpHc197b44nU,1017
 agent/providers/omni/utils.py,sha256=JqSye1bEp4wxhUgmaMyZi172fTlgXtygJ7XlnvKdUtE,6337
 agent/providers/omni/visualization.py,sha256=N3qVQLxYmia3iSVC5oCt5YRlMPuVfylCOyB99R33u8U,3924
+agent/telemetry.py,sha256=pVGxbj0ewnvq4EGj28CydN4a1iOfvZR_XKL3vIOqhOM,390
 agent/types/__init__.py,sha256=61UFJT-w0CT4YRn0LiTx4A7fsMdVQjlXO9vnmbI1A7Y,604
 agent/types/base.py,sha256=Iy_Q2DIBMLtwWdLyfvHw_6E2ltYu3bIv8GUNy3LYkGs,1133
 agent/types/messages.py,sha256=4-hwtxeAhto90_EZpHFducddtsHUsHauvXzYrpKG4RE,953
 agent/types/tools.py,sha256=Jes2CFCFqC727WWHbO-sG7V03rBHnQe5X7Oi9ZkuScI,877
-cua_agent-0.1.2.dist-info/METADATA,sha256=bXSToJpS_e5KRzyRELUzCuOkozsDUD29pBMj3DKzF7U,1890
-cua_agent-0.1.2.dist-info/WHEEL,sha256=thaaA2w1JzcGC48WYufAs8nrYZjJm8LqNfnXFOFyCC4,90
-cua_agent-0.1.2.dist-info/entry_points.txt,sha256=6OYgBcLyFCUgeqLgnvMyOJxPCWzgy7se4rLPKtNonMs,34
-cua_agent-0.1.2.dist-info/RECORD,,
+cua_agent-0.1.4.dist-info/METADATA,sha256=9AW-hOrH5hmJ3Al9sFgoYV-GUCdhfoRzYYqYAMw2bpA,4528
+cua_agent-0.1.4.dist-info/WHEEL,sha256=thaaA2w1JzcGC48WYufAs8nrYZjJm8LqNfnXFOFyCC4,90
+cua_agent-0.1.4.dist-info/entry_points.txt,sha256=6OYgBcLyFCUgeqLgnvMyOJxPCWzgy7se4rLPKtNonMs,34
+cua_agent-0.1.4.dist-info/RECORD,,

cua_agent-0.1.2.dist-info/METADATA

@@ -1,44 +0,0 @@
-Metadata-Version: 2.1
-Name: cua-agent
-Version: 0.1.2
-Summary: CUA (Computer Use) Agent for AI-driven computer interaction
-Author-Email: TryCua <gh@trycua.com>
-Requires-Python: <3.13,>=3.10
-Requires-Dist: httpx<0.29.0,>=0.27.0
-Requires-Dist: aiohttp<4.0.0,>=3.9.3
-Requires-Dist: asyncio
-Requires-Dist: anyio<5.0.0,>=4.4.1
-Requires-Dist: typing-extensions<5.0.0,>=4.12.2
-Requires-Dist: pydantic<3.0.0,>=2.6.4
-Requires-Dist: rich<14.0.0,>=13.7.1
-Requires-Dist: python-dotenv<2.0.0,>=1.0.1
-Requires-Dist: cua-computer<0.2.0,>=0.1.0
-Requires-Dist: certifi>=2024.2.2
-Provides-Extra: anthropic
-Requires-Dist: anthropic>=0.49.0; extra == "anthropic"
-Requires-Dist: boto3<2.0.0,>=1.35.81; extra == "anthropic"
-Provides-Extra: som
-Requires-Dist: torch>=2.2.1; extra == "som"
-Requires-Dist: torchvision>=0.17.1; extra == "som"
-Requires-Dist: ultralytics>=8.0.0; extra == "som"
-Requires-Dist: transformers>=4.38.2; extra == "som"
-Requires-Dist: cua-som<0.2.0,>=0.1.0; extra == "som"
-Requires-Dist: anthropic<0.47.0,>=0.46.0; extra == "som"
-Requires-Dist: boto3<2.0.0,>=1.35.81; extra == "som"
-Requires-Dist: openai<2.0.0,>=1.14.0; extra == "som"
-Requires-Dist: groq<0.5.0,>=0.4.0; extra == "som"
-Requires-Dist: dashscope<2.0.0,>=1.13.0; extra == "som"
-Requires-Dist: requests<3.0.0,>=2.31.0; extra == "som"
-Provides-Extra: all
-Requires-Dist: torch>=2.2.1; extra == "all"
-Requires-Dist: torchvision>=0.17.1; extra == "all"
-Requires-Dist: ultralytics>=8.0.0; extra == "all"
-Requires-Dist: transformers>=4.38.2; extra == "all"
-Requires-Dist: cua-som<0.2.0,>=0.1.0; extra == "all"
-Requires-Dist: anthropic<0.47.0,>=0.46.0; extra == "all"
-Requires-Dist: boto3<2.0.0,>=1.35.81; extra == "all"
-Requires-Dist: openai<2.0.0,>=1.14.0; extra == "all"
-Requires-Dist: groq<0.5.0,>=0.4.0; extra == "all"
-Requires-Dist: dashscope<2.0.0,>=1.13.0; extra == "all"
-Requires-Dist: requests<3.0.0,>=2.31.0; extra == "all"
-