UCloud ModelVerse 图片 API 用法

UCloud ModelVerse 图片 API 用法 接口： POST https://api.umodelverse.ai/v1/images/generations 请求头： Authorization: Bearer YOUR_API_KEY Content-Type: application/json 推荐参数： { "model": "gpt-image-1.5", "prompt": "英文图片提示词", "n": 1, "size": "1536x1024", "quality": "high" } 注意：这次实测 UCloud 不接受 response_format，加了会报： Invalid param: Unknown parameter: 'response_format' 所以不要传 response_format。返回里可能是 b64_json，也可能是 url，脚本要两种都兼容。 标准 Python 调用模板 import base64 import json import os import urllib.request import urllib.error from pathlib import Path API_URL = "https://api.umodelverse.ai/v1/images/generations" MODEL = "gpt-image-1.5" api_key = os.environ["MODELVERSE_API_KEY"] prompt = """ Create a realistic B2B RF antenna industry blog image. No readable text, no labels, no numbers, no Chinese characters, no watermark, no logo, no brand names. Professional engineering photography style. """ payload = { "model": MODEL, "prompt": prompt, "n": 1, "size": "1536x1024", "quality": "high" } req = urllib.request.Request( API_URL, data=json.dumps(payload).encode("utf-8"), headers={ "Authorization": f"Bearer {api_key}", "Content-Type": "application/json", }, method="POST", ) try: with urllib.request.urlopen(req, timeout=300) as resp: body = resp.read().decode("utf-8") except urllib.error.HTTPError as e: err = e.read().decode("utf-8", errors="replace") raise RuntimeError(f"UCloud API {e.code}: {err}") from e data = json.loads(body) image = data["data"][0] if image.get("b64_json"): image_bytes = base64.b64decode(image["b64_json"]) elif image.get("url"): with urllib.request.urlopen(image["url"], timeout=300) as img_resp: image_bytes = img_resp.read() else: raise RuntimeError(f"Unsupported response: {body[:1000]}") Path("output.png").write_bytes(image_bytes) 终端使用方式 不要把 key 写进脚本或日志，只临时放环境变量： export MODELVERSE_API_KEY='你的UCloud API Key' python3 generate_images.py