How to run Claude computer use demo on macOS

README.md

Note

It is necessary to give Terminal (or iTerm or whatever you use) the permission to control the computer. This can be done in System Settings ➔ Privacy & Security ➔ Accessibility.

Guide

• Install cliclick for mouse & keyboard emulation
  • brew install cliclick
• Clone Anthropic quickstart repo
  • git clone https://github.com/anthropics/anthropic-quickstarts.git
• cd computer-use-demo
• Replace computer-use-demo/computer_use_demo/tools/computer.py with the modified version below
• ./setup.sh
• source .venv/bin/activate
• Export enviroment varaiables
  • export ANTHROPIC_API_KEY=sk-xxxxxx // replace with your api key
  • export WIDTH=1512 // replace with your screen resolution
  • export HEIGHT=982 // replace with your screen resolution
• Start the app
  • python -m streamlit run computer_use_demo/streamlit.py

computer.py

import asyncio
import base64
import os
import shlex
import shutil
from enum import StrEnum
from pathlib import Path
from typing import Literal, TypedDict
from uuid import uuid4

from anthropic.types.beta import BetaToolComputerUse20241022Param

from .base import BaseAnthropicTool, ToolError, ToolResult
from .run import run

OUTPUT_DIR = "/tmp/outputs"

TYPING_DELAY_MS = 12
TYPING_GROUP_SIZE = 50

IMAGE_MAX_WIDTH = 1200

Action = Literal[
    "key",
    "type",
    "mouse_move",
    "left_click",
    "left_click_drag",
    "right_click",
    "middle_click",
    "double_click",
    "screenshot",
    "cursor_position",
]


class Resolution(TypedDict):
    width: int
    height: int


class ScalingSource(StrEnum):
    COMPUTER = "computer"
    API = "api"


class ComputerToolOptions(TypedDict):
    display_height_px: int
    display_width_px: int
    display_number: int | None


def chunks(s: str, chunk_size: int) -> list[str]:
    return [s[i : i + chunk_size] for i in range(0, len(s), chunk_size)]


class ComputerTool(BaseAnthropicTool):
    """
    A tool that allows the agent to interact with the screen, keyboard, and mouse of the current computer.
    The tool parameters are defined by Anthropic and are not editable.
    """

    name: Literal["computer"] = "computer"
    api_type: Literal["computer_20241022"] = "computer_20241022"
    width: int
    height: int
    display_num: int | None

    _screenshot_delay = 2.0
    _scaling_enabled = True

    @property
    def options(self) -> ComputerToolOptions:
        width, height = self.scale_coordinates(
            ScalingSource.COMPUTER, self.width, self.height
        )
        return {
            "display_width_px": width,
            "display_height_px": height,
            "display_number": self.display_num,
        }

    def to_params(self) -> BetaToolComputerUse20241022Param:
        return {"name": self.name, "type": self.api_type, **self.options}

    def __init__(self):
        super().__init__()

        self.width = int(os.getenv("WIDTH") or 0)
        self.height = int(os.getenv("HEIGHT") or 0)
        assert self.width and self.height, "WIDTH, HEIGHT must be set"
        if (display_num := os.getenv("DISPLAY_NUM")) is not None:
            self.display_num = int(display_num)
            self._display_prefix = f"DISPLAY=:{self.display_num} "
        else:
            self.display_num = None
            self._display_prefix = ""

        self.xdotool = f"{self._display_prefix}xdotool"

    async def __call__(
        self,
        *,
        action: Action,
        text: str | None = None,
        coordinate: tuple[int, int] | None = None,
        **kwargs,
    ):
        if action in ("mouse_move", "left_click_drag"):
            if coordinate is None:
                raise ToolError(f"coordinate is required for {action}")
            if text is not None:
                raise ToolError(f"text is not accepted for {action}")
            if not isinstance(coordinate, list) or len(coordinate) != 2:
                raise ToolError(f"{coordinate} must be a tuple of length 2")
            if not all(isinstance(i, int) and i >= 0 for i in coordinate):
                raise ToolError(f"{coordinate} must be a tuple of non-negative ints")

            x, y = self.scale_coordinates(
                ScalingSource.API, coordinate[0], coordinate[1]
            )

            if action == "mouse_move":
                return await self.shell(f"cliclick m:{x},{y}")
            elif action == "left_click_drag":
                return await self.shell(f"cliclick dd:{x},{y}")

        if action in ("key", "type"):
            if text is None:
                raise ToolError(f"text is required for {action}")
            if coordinate is not None:
                raise ToolError(f"coordinate is not accepted for {action}")
            if not isinstance(text, str):
                raise ToolError(output=f"{text} must be a string")

            if action == "key":
                return await self.shell(f"cliclick kp:{text.lower()}")
            elif action == "type":
                results: list[ToolResult] = []
                for chunk in chunks(text, TYPING_GROUP_SIZE):
                    cmd = f"cliclick t:'{chunk}'"
                    results.append(await self.shell(cmd, take_screenshot=False))
                screenshot_base64 = (await self.screenshot()).base64_image
                return ToolResult(
                    output="".join(result.output or "" for result in results),
                    error="".join(result.error or "" for result in results),
                    base64_image=screenshot_base64,
                )

        if action in (
            "left_click",
            "right_click",
            "double_click",
            "middle_click",
            "screenshot",
            "cursor_position",
        ):
            if text is not None:
                raise ToolError(f"text is not accepted for {action}")
            if coordinate is not None:
                raise ToolError(f"coordinate is not accepted for {action}")

            if action == "screenshot":
                return await self.screenshot()
            elif action == "cursor_position":
                result = await self.shell(
                    f"cliclick p",
                    take_screenshot=False,
                )
                output = result.output or ""
                x, y = self.scale_coordinates(
                    ScalingSource.COMPUTER,
                    int(output.split("X=")[1].split("\n")[0]),
                    int(output.split("Y=")[1].split("\n")[0]),
                )
                return result.replace(output=f"X={x},Y={y}")
            else:
                click_arg = {
                    "left_click": "c:.",
                    "right_click": "rc:.",
                    "middle_click": "tc:.",
                    "double_click": "dc:.",
                }[action]
                return await self.shell(f"cliclick {click_arg}")

        raise ToolError(f"Invalid action: {action}")

    async def screenshot(self):
        """Take a screenshot of the current screen and return the base64 encoded image."""
        output_dir = Path(OUTPUT_DIR)
        output_dir.mkdir(parents=True, exist_ok=True)
        path = output_dir / f"screenshot_{uuid4().hex}.png"

        screenshot_cmd = f"{self._display_prefix}screencapture {path}"

        result = await self.shell(screenshot_cmd, take_screenshot=False)
        if self._scaling_enabled:
            x, y = self.scale_coordinates(
                ScalingSource.COMPUTER, self.width, self.height
            )
            await self.shell(
                f"convert {path} -resize {x}x{y}! {path}", take_screenshot=False
            )

        if path.exists():
            return result.replace(
                base64_image=base64.b64encode(path.read_bytes()).decode()
            )
        raise ToolError(f"Failed to take screenshot: {result.error}")

    async def shell(self, command: str, take_screenshot=True) -> ToolResult:
        """Run a shell command and return the output, error, and optionally a screenshot."""
        _, stdout, stderr = await run(command)
        base64_image = None

        if take_screenshot:
            # delay to let things settle before taking a screenshot
            await asyncio.sleep(self._screenshot_delay)
            base64_image = (await self.screenshot()).base64_image

        return ToolResult(output=stdout, error=stderr, base64_image=base64_image)

    def scale_coordinates(self, source: ScalingSource, x: int, y: int):
        """Scale coordinates to a target maximum resolution."""
        if not self._scaling_enabled:
            return x, y
        ratio = self.width / self.height
        target_dimension = { "width": IMAGE_MAX_WIDTH, "height": IMAGE_MAX_WIDTH / ratio }
        # should be less than 1
        x_scaling_factor = target_dimension["width"] / self.width
        y_scaling_factor = target_dimension["height"] / self.height
        if source == ScalingSource.API:
            if x > self.width or y > self.height:
                raise ToolError(f"Coordinates {x}, {y} are out of bounds")
            # scale up
            return round(x / x_scaling_factor), round(y / y_scaling_factor)
        # scale down
        return round(x * x_scaling_factor), round(y * y_scaling_factor)

@moo1o Did you forget to replace computer.py?

Sorry, I changed it wrong.

@wong2 hello The execution was successful. However, since the code is written based on Ubuntu, claude is trying to take a screenshot using the scrot package. Also, he is trying to use bash to install the scrot package. Does it work normally?

Is the agent ultimately able to install the screen capture tool?
Because the system prompt usually says it is running on Linux.

@valdo404 screencapture is a built-in command in macOS

I think xdotool is available on MacOS too? Does it not work as expected?

Now macOS no longer comes with X11, and even if you install XQuartz, it still won't work properly.

anyone tried a version of this for windows?