GPT4Free TypeScript Version: Providing a free OpenAI GPT-4 API!

GPT4Free TypeScript Version is a project that aims to provide a free GPT-4 API, a powerful natural language processing model developed by OpenAI. The project is based on TypeScript, a programming language that extends JavaScript with static types and other features. The project uses TensorFlow.js, a library for machine learning in JavaScript, to implement the GPT-4 model and run it in the browser or on Node.js.

To get started, A file named .env is required for this project.

This step is essential for all methods of operation.

http_proxy=http://host:port
rapid_api_key=xxxxxxxxxx
EMAIL_TYPE=temp-email44
DEBUG=0
POOL_SIZE=1

• http_proxy: If you have trouble reaching the desired site, you may need to set up a proxy server.
• rapid_api_key: This is a required configuration for using the forefront api. The apikey allows you to receive registration emails. get api key here
• EMAIL_TYPE: temporary email types include temp-email temp-email44 tempmail-lol
  • temp-email: The maximum number of requests per day is 100. To increase this limit, you need to link a credit card to your account. This service is very reliable!
  • temp-email44: limited 100req/days! Stable!
  • tempmail-lol: limited to 25request/5min. Not Stable.
• DEBUG: Valid when use forefront Set it to 1 when running locally. show reverse process
• POOL_SIZE: forefront concurrency size. Keep set=1 until you run it successfully!!! You can engage in {POOL_SIZE} conversations concurrently. A higher pool size enables more parallel conversations, but it also consumes more RAM.

Run it locally 🖥️

# install module
yarn
# start server
yarn start

Run with docker🐳(Recommended!)

docker run -p 3000:3000 --env-file .env xiangsx/gpt4free-ts:latest

Deploy with docker-compose 🎭

To run the application with docker, you need to create a .env file first. Then, follow the instructions in the “Run with docker” section.

deployment:

docker-compose up --build -d

Request Parameters📝

• prompt: your question. It can be a string or jsonstr.
  • example jsonstr:[{"role":"user","content":"你好\n"},{"role":"assistant","content":"你好！有什么我可以帮助你的吗？"},{"role":"user","content":"你是谁"}]
  • example string: 你是谁
• model: default gpt3.5-turbo. model include:gpt4 gpt3.5-turbo
• site: default you. target site, include forefront you mcbbs

Response Parameters 🔙

Response when chat ends (/ask):

interface ChatResponse {
    content: string;
    error?: string;
}

Response with stream like(/ask/stream):

event: message
data: {"content":"I"}

event: done
data: {"content":"'m"}

event: error
data: {"error":"some thind wrong"}

Example💡

1. request to site you with history

req:

127.0.0.1:3000/ask?site=you&prompt=[{“role”:”user”,”content”:”hello”},{“role”:”assistant”,”content”:”Hi there! How can I assist you today?”},{“role”:”user”,”content”:”who are you”}]

res:

{
  "content": "Hi there! How can I assist you today?"
}

This GitHub repository contains APIs from various sites without their authorization or affiliation. This project is for learning purposes only. This is a personal hobby project. Sites can contact developer to request better security or removal of their site from this repository.