[POST]
summary = "分析图片内容"
docstring = "发送图片（base64或路径）到豆包视觉模型，返回描述"
params = [
    "image_base64: str",
    "prompt: str"
]
response_model = { result = { type = "str", required = true }, error = { type = "str", required = false, default = "" } }