Computer Use Beta

import base64
from GlomaxGPT import GlomaxGPT
from playwright.sync_api import sync_playwright

client = GlomaxGPT()

def take_screenshot(page):
    """Take a screenshot and return as base64."""
    screenshot_bytes = page.screenshot()
    return base64.b64encode(screenshot_bytes).decode("utf-8")

def execute_action(page, action):
    """Execute a computer action on the page."""
    if action.type == "click":
        page.mouse.click(action.coordinate[0], action.coordinate[1])
    elif action.type == "type":
        page.keyboard.type(action.text)
    elif action.type == "scroll":
        page.mouse.wheel(action.coordinate[0], action.coordinate[1])
    elif action.type == "key":
        page.keyboard.press(action.key)
    elif action.type == "screenshot":
        pass  # Just take a new screenshot

with sync_playwright() as p:
    browser = p.chromium.launch(headless=True)
    page = browser.new_page(viewport={"width": 1280, "height": 720})
    page.goto("https://example.com")

    task = "Navigate to the About page and extract the company founding year."
    screenshot_b64 = take_screenshot(page)

    messages = [{
        "role": "user",
        "content": [
            {"type": "text", "text": task},
            {
                "type": "input_image",
                "image_url": f"data:image/png;base64,{screenshot_b64}"
            }
        ]
    }]

    while True:
        response = client.responses.create(
            model="glomaxgpt-ultra",
            input=messages,
            tools=[{
                "type": "computer_use_preview",
                "display_width": 1280,
                "display_height": 720,
                "environment": "browser"
            }],
            truncation="auto"
        )

        # Check for computer use actions
        computer_calls = [
            item for item in response.output
            if item.type == "computer_call"
        ]

        if not computer_calls:
            # No more actions — task is complete
            print(response.output_text)
            break

        for call in computer_calls:
            execute_action(page, call.action)

        # Take new screenshot and continue loop
        new_screenshot = take_screenshot(page)
        messages.append({
            "role": "tool",
            "tool_call_id": computer_calls[-1].id,
            "content": [{
                "type": "input_image",
                "image_url": f"data:image/png;base64,{new_screenshot}"
            }]
        })

    browser.close()