openai proxy Download - openai proxy Source code download

openai proxy

Other source code

v0.3.15

Download

Hi, there

Please check the API documentation https://openai-proxy.apifox.cn

openai-proxy is a solution for Chinese users. The problems that can help you solve are:

✅ OpenAI API request timeout
✅ OpenAI API does not support querying cost consumption
✅ Claude API is still waiting for the list
✅ Google Bard does not provide API
. . .

OpenAI API request timeout

Now you just need to replace https://api.openai.com with https://openai.aihey.cc/openai to eat

 # openai api
curl https://api.openai.com/v1/chat/completions 
  -H " Content-Type: application/json " 
  -H " Authorization: Bearer $OPENAI_API_KEY " 
  -d ' {
    "model": "gpt-3.5-turbo",
    "messages": [{"role": "user", "content": "Hello!"}]
  } '

# 替换后
curl https://openai.aihey.cc/openai/v1/chat/completions 
  -H " Content-Type: application/json " 
  -H " Authorization: Bearer $OPENAI_API_KEY " 
  -d ' {
    "model": "gpt-3.5-turbo",
    "messages": [{"role": "user", "content": "Hello!"}]
  } '

OpenAI API does not support querying cost consumption

Since April 2023, https://api.openai.com/dashboard/billing/credit_grants can only be requested through the session id generated by the web login, and the key ( sk-***** ) method will be invalid. However, you can request it in two ways .

Method 1: Use `openai-proxy`

shortcoming:

There may be differences between the data displayed on the web page (I don’t know why this is. Friends who are familiar with it are welcome to point it out in ISSUES)

 # 指定统计开始日期和结束日期
curl https://openai.aihey.cc/openai/billing/credit_grants ? start_date=2022-05-10 & end_date=2023-05-10 
  -H " Content-Type: application/json " 
  -H " Authorization: Bearer sk-*********** "

# 不指定参数，默认统计最近 90 天的数据
curl https://openai.aihey.cc/openai/billing/credit_grants 
  -H " Content-Type: application/json " 
  -H " Authorization: Bearer sk-********** "

Method 2: Use `sensitive_id` generated by the web page

shortcoming:

You need to go to the webpage to get it for the first time (press F12 and refresh the webpage)
It may be invalid at any time (but at present, as long as the web page does not log out, sensitive_id will be relatively long)

Method 3: Use Docker

docker run -itd --name openai-proxy -p 13000:3000  unickcheng/openai-proxy

# 查看服务
docker ps -a

If you are familiar with docker compose, you can refer to the docker-compose.yml file and use docker compose up -d command to start the service

Some questions

Will the openai api key be stolen (key point)

No, but don't let go of your vigilance . There are also many excellent products of similar open source on the market, and they will not maliciously steal the user's key. Unfortunately, this does not mean everything. Therefore, when using third-party platforms products, try to use a regularly updated key .

If you have the conditions, it is recommended to refer to the previous content to deploy it yourself. If you encounter any problems, you can ask questions in Issues.

Why create this project

Because I wanted to get started with Next.js, I had this project.

In addition, here is just a solution, which has not improved much for API request speed. You need to provide additional domain names by yourself, which seems to be troublesome.

If you are a developer who loves tossing, you may also try Nginx, Cloudflare and other solutions
If you want to have a solution to use it directly, please feel free to be proud of yourself