chatgpt 4V 识图功能

1.获取图片的sig和file_id

2e0edc6e489ed13a3f32f0dd87527d77.jpg是本地图片的名字

头部认证信息自己F12 抓取

复制代码
1.获取图片的sig

https://chat.openai.com/backend-api/files

Authorization:Bearer eyJhbGc****************5V-lztYwLb9hr6LP7g
Cookie:  *************************D


{"file_name":"2e0edc6e489ed13a3f32f0dd87527d77.jpg","file_size":51010,"use_case":"multimodal"}

返回{
    "status": "success",
    "upload_url": "https://fileserviceuploadsperm.blob.core.windows.net/files/file-KVRbTP1Xy0NP1WKKZRDKSPL1?se=2023-10-16T11%3A24%3A03Z&sp=c&sv=2021-08-06&sr=b&sig=oXzcBB7Q8HWyZr6JUSbuUYtgwgOWVdia7EiO8ALBe%2Bw%3D",
    "file_id": "file-KVRbTP1Xy0NP1WKKZRDKSPL1"
}

2.上传二进制图片

复制代码
https://fileserviceuploadsperm.blob.core.windows.net/files/file-KVRbTP1Xy0NP1WKKZRDKSPL1?se=2023-10-16T11%3A24%3A03Z&sp=c&sv=2021-08-06&sr=b&sig=oXzcBB7Q8HWyZr6JUSbuUYtgwgOWVdia7EiO8ALBe%2Bw%3D


Accept: application/json, text/plain, */*
Accept-Encoding: gzip, deflate, br
Accept-Language: zh-CN,zh;q=0.9,en;q=0.8,en-GB;q=0.7,en-US;q=0.6
Connection: keep-alive

Content-Type: image/jpeg
Host: fileserviceuploadsperm.blob.core.windows.net
Origin: https://chat.openai.com
Sec-Fetch-Dest: empty
Sec-Fetch-Mode: cors
Sec-Fetch-Site: cross-site
User-Agent: Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/117.0.0.0 Safari/537.36 Edg/117.0.2045.60
sec-ch-ua: "Microsoft Edge";v="117", "Not;A=Brand";v="8", "Chromium";v="117"
sec-ch-ua-mobile: ?0
sec-ch-ua-platform: "Windows"
x-ms-blob-type: BlockBlob
x-ms-version: 2020-04-08
无返回

3.put图片

复制代码
https://fileserviceuploadsperm.blob.core.windows.net/files/file-KVRbTP1Xy0NP1WKKZRDKSPL1?se=2023-10-16T11%3A24%3A03Z&sp=c&sv=2021-08-06&sr=b&sig=oXzcBB7Q8HWyZr6JUSbuUYtgwgOWVdia7EiO8ALBe%2Bw%3D
Accept: application/json, text/plain, */*
Accept-Encoding: gzip, deflate, br
Accept-Language: zh-CN,zh;q=0.9,en;q=0.8,en-GB;q=0.7,en-US;q=0.6
Connection: keep-alive
Content-Length: 51010
Content-Type: image/jpeg
Host: fileserviceuploadsperm.blob.core.windows.net
Origin: https://chat.openai.com
Sec-Fetch-Dest: empty
Sec-Fetch-Mode: cors
Sec-Fetch-Site: cross-site
User-Agent: Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/117.0.0.0 Safari/537.36 Edg/117.0.2045.60
sec-ch-ua: "Microsoft Edge";v="117", "Not;A=Brand";v="8", "Chromium";v="117"
sec-ch-ua-mobile: ?0
sec-ch-ua-platform: "Windows"
x-ms-blob-type: BlockBlob
x-ms-version: 2020-04-08

form-data
本地图片地址

4.上传二进制图片

复制代码
4.上传二进制图片
https://chat.openai.com/backend-api/files/file-KVRbTP1Xy0NP1WKKZRDKSPL1/uploaded

Authorization:Bearer eyJ*****tYwLb9hr6LP7g
Cookie:  *HS%2BvXw%3D

提交空
{}

返回
{
    "status": "success",
    "download_url": "https://fileserviceuploadsperm.blob.core.windows.net/files/file-KVRbTP1Xy0NP1WKKZRDKSPL1?se=2023-10-16T11%3A34%3A11Z&sp=r&sv=2021-08-06&sr=b&rscd=attachment%3B%20filename%3D2e0edc6e489ed13a3f32f0dd87527d77.jpg&sig=LJbEvMeV6KtlkTi6Y1udDd%2BK3YSbgfUWnsGCEi5rGSs%3D",
    "metadata": null
}

5.和chatgpt问题绑定,提问

复制代码
https://chat.openai.com/backend-api/conversation
Authorization:Bearer eyJhbGciOiJSU******%3D


{
	"action": "next",
	"messages": [{
		"id": "aaa27270-f62e-454a-962c-f62794152450",
		"author": {
			"role": "user"
		},
		"content": {
			"content_type": "multimodal_text",
			"parts": [{
				"asset_pointer": "file-service://file-KVRbTP1Xy0NP1WKKZRDKSPL1",
				"size_bytes": 239505,
				"width": 1706,
				"height": 1280
			}, "再看看"]
		},
		"metadata": {}
	}],
	"conversation_id": "41181b17-b71d-4748-8db5-3a80d7a27a33",
	"parent_message_id": "49bd950f-372f-45f6-9f72-b6f27fa04373",
	"model": "gpt-4",
	"timezone_offset_min": -480,
	"suggestions": [],
	"history_and_training_disabled": false,

	"force_paragen": false
}

返回答案
相关推荐
余俊晖1 小时前
多模态文档解析新进展:多模态OCR解析文档中的任意内容实现方案
人工智能·自然语言处理·多模态
余俊晖1 小时前
多模态文档解析最新开源进展:2B参数FireRed-OCR模型方法、数据
人工智能·自然语言处理·ocr·多模态
智算菩萨3 小时前
ChatGPT在非洲主要国家教育中的应用:效益、接受度与伦理挑战——基于2022-2024年文献的系统综述精读
论文阅读·人工智能·gpt·深度学习·ai·chatgpt·论文笔记
柯儿的天空21 小时前
【OpenClaw 全面解析:从零到精通】第 006 篇:OpenClaw 在 Windows/WSL2 上的安装与部署实战
人工智能·windows·语言模型·chatgpt·ai作画
Agent产品评测局1 天前
中国龙虾ai软件有哪些选择?2026自动化选型指南
运维·人工智能·ai·chatgpt·自动化
人道领域1 天前
2026年Q1大模型深度复盘:OpenAI,Gemini2.0,字节跳动,与“多模态Agent”元年
人工智能·ai·google·chatgpt·gemini
柯儿的天空1 天前
【OpenClaw 全面解析:从零到精通】第 010 篇:OpenClaw多渠道接入:WhatsApp、Telegram、飞书等
人工智能·chatgpt·ai作画·aigc·飞书·ai编程·ai写作
小锋学长生活大爆炸2 天前
【工具】无需Token!WebAI2API将网页AI转为API使用
人工智能·深度学习·chatgpt·openclaw
_张一凡2 天前
【大语言模型学习】一文详解阿里Qwen3大模型以及全参量微调入门实战教程(代码完整)
llm·aigc·大语言模型·多模态·qwen3·大语言模型微调·全参量微调
大傻^2 天前
Spring AI 2.0 企业级 RAG 架构:混合检索、重排序与多模态知识库
人工智能·spring·架构·多模态·rag·混合检索·重排序