from openai import OpenAI from playwright.sync_api import sync_playwright import base64 client = OpenAI( api_key="your-key", base_url="https://api.ai.cc/v1") def capture(page): return base64.b64encode(page.screenshot()).decode() with sync_playwright() as p: browser = p.chromium.launch(headless=False) page = browser.new_page() page.goto("https://example.com") response = client.responses.create( model="gpt-5.4", tools=[{"type": "computer"}], input="Search latest AI news") while True: call = next((x for x in response.output if x.type="computer_call"), None) if not call: break for act in call.actions: if act.type="click": page.mouse.click(act.x, act.y) elif act.type="type": page.keyboard.type(act.text) response = client.responses.create( model="gpt-5.4", previous_response_id=response.id, tools=[{"type":"computer"}], input=[{ "type":"computer_call_output", "call_id":call.call_id, "output":{ "type":"computer_screenshot", "image_url":"data:image/png;base64,"+capture(page) } }] )