chatgpt4的免费api，还支持以下模型：gpt-4-all,gpt-3.5-turbo,gpt-3.5-turbo-16k

<div align="center"> # GPT4Free TypeScript Version 🆓 ###### Providing a free OpenAI GPT-4 API! English | [中文](README_zh.md) | [日本語](README_ja.md) [](https://discord.gg/cYUU8mCDMd) <p>You can join our discord: <a href="https://discord.gg/cYUU8mCDMd">discord.gg/gptgod<a> for further updates. <a href="https://discord.gg/cYUU8mCDMd"><img align="center" alt="gpt4free Discord" width="22px" src="https://raw.githubusercontent.com/peterthehan/peterthehan/master/assets/discord.svg" /></a></p> </div> ## 🆓 Free OpenAI Key Base URL: `https://api.gptgod.online` Api Key: `sk-OsMMq65tXdfOIlTUYtocSL7NCsmA7CerN77OkEv29dODg1EA` Support Models: `gpt-4-all`,`gpt-3.5-turbo`,`gpt-3.5-turbo-16k`,`net-gpt-3.5-turbo`,`net-gpt-3.5-turbo-16k`,`claude-1-100k`,`google-palm`,`llama-2-70b`,`llama-2-13b`,`llama-2-7b`,`code-llama-34b`,`code-llama-13b`,`code-llama-7b`,`qwen-72b`,`stable-diffusion`,`mixtral-8x7b`,`mistral-medium` ## 🚩 Reverse target I suggest you fork this project first. Some websites may go offline at any time. Still striving to keep updating. Have implemented models here: If you do not want your website to appear here, please raise an issue and I will remove it immediately. Unfortunately, most of the sites here are no longer available. ***Update At 2023-09-10*** | Site | Models | |----------|---------------------------------------------------| | you | gpt-3.5-turbo | | phind | net-gpt-3.5-turbo | | forefront| gpt-3.5-turbo, claude | | mcbbs | gpt-3.5-turbo, gpt-3.5-turbo-16k | | chatdemo | gpt-3.5-turbo, gpt-3.5-turbo-16k | | vita | gpt-3.5-turbo | | skailar | gpt-4 | | fakeopen | gpt-3.5-turbo, gpt-3.5-turbo-16k, gpt-4 | | easychat | gpt-4 | | better | gpt-3.5-turbo, gpt-3.5-turbo-16k, gpt-4 | | pweb | gpt-3.5-turbo, gpt-3.5-turbo-16k | | bai | gpt-3.5-turbo | | gra | gpt-3.5-turbo, gpt-3.5-turbo-16k | | magic | gpt-3.5-turbo, gpt-4, claude-instance, claude, claude-100k | | chim | gpt-3.5-turbo, gpt-3.5-turbo-16k, gpt-4 | | ram | gpt-3.5-turbo-16k | | chur | gpt-3.5-turbo, gpt-3.5-turbo-16k | | xun | gpt-3.5-turbo, gpt-3.5-turbo-16k | | vvm | gpt-3.5-turbo, gpt-3.5-turbo-16k, gpt-4 | | poef | | | claude | claude-2-100k | | cursor | gpt-3.5-turbo, gpt-4 | | chatbase | gpt-3.5-turbo | | ails | gpt-3.5-turbo | | sincode | gpt-3.5-turbo, gpt-4 | | openai | too much | | jasper | gpt-3.5-turbo, gpt-4 | | pap | | | acytoo | gpt-3.5-turbo | | google | search | | www | url | | ddg | search | ## 🏃‍♂️ Run First of all, you should create file `.env`. > ***All operation methods require this step.*** ```env http_proxy=http://host:port rapid_api_key=xxxxxxxxxx EMAIL_TYPE=temp-email44 DEBUG=0 POOL_SIZE=0 PHIND_POOL_SIZE=0 ``` - `http_proxy`: config your proxy if you can not access target website directly; If you dont need proxy, delete this line; - `forefront` use env(this site has been removed): - `rapid_api_key`: you should config this if you use forefront api, this apikey is used for receive register email, get api key here - `EMAIL_TYPE`: temp email type includes `temp-email` `temp-email44` `tempmail-lol` - [temp-email](https://rapidapi.com/Privatix/api/temp-mail): soft limit 100req/days, if over use money, need bind credit card! Very Stable! - [temp-email44](https://rapidapi.com/calvinloveland335703-0p6BxLYIH8f/api/temp-mail44): hard limit 100req/days! Stable! - [tempmail-lol](): nothing need, limit 25request/5min. Not Stable. - `DEBUG`: Valid when use `forefront` You can set =1 when you run local. show reverse process - `POOL_SIZE`: `forefront` concurrency size. Keep set=1 until you run it successfully!!! You can engage in {POOL_SIZE} conversations concurrently. More pool size, More conversation can be done simultaneously, But use more RAM - `phind` use env: - `PHIND_POOL_SIZE`: `phind` concurrency size.You can engage in {POOL_SIZE} conversations concurrently. More pool size, More conversation can be done simultaneously, But use more RAM ### Run local 🖥️ ```shell # install module yarn # start server yarn start ``` ### Run with docker(Suggest!) 🐳 ``` docker run -p 3000:3000 --env-file .env xiangsx/gpt4free-ts:latest ``` ### Deploy with docker-compose 🎭 first, you should create file .env; Follow step "Run with docker deploy ``` docker-compose up --build -d ``` ### Clash+one-api+gpt4free-ts Start with one command 😮 [gpt4free-ts-deploy](https://github.com/xiangsx/gpt4free-ts-deploy) ## 🚀 Let's Use GPT4 > Find supports model and site http://127.0.0.1:3000/supports [GET] > The same as openai http://127.0.0.1:3000/:site/v1/chat/completions [POST] > The same as openai http://127.0.0.1:3000/v1/chat/completions?site=xxx [POST] > Return when chat complete http://127.0.0.1:3000/ask?prompt=***&model=***&site=*** [POST/GET] > Return with eventstream http://127.0.0.1:3000/ask/stream?prompt=***&model=***&site=*** [POST/GET] ### Request Params 📝 - `prompt`: your question. It can be a `string` or `jsonstr`. - example `jsonstr`:`[{"role":"user","content":"hello\n"},{"role":"assistant","content":"Hi there! How can I assist you today?"},{"role":"user","content":"who are you"}]` - example `string`: `who are you` - `model`: default `gpt3.5-turbo`. model include:`gpt4` `gpt3.5-turbo` `net-gpt3.5-turbo` `gpt-3.5-turbo-16k` - `site`: default `you`. target site, include `fakeopen` `better` `forefront` `you` `chatdemo` `phind` `vita` ### Site Support Model 🧩 query supports site and models with api `127.0.0.1:3000/supports` ```json [ { "site": "you", "models": [ "gpt-3.5-turbo" ] }, ... ] ``` ### Response Params 🔙 Response when chat end(/ask): ```typescript interface ChatResponse { content: string; error?: string; } ``` Response with stream like, Suggest!!(/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: ```json { "content": "Hi there! How can I assist you today?" } ``` [127.0.0.1:3000/ask?site=you&prompt=[{"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"}]]() 2. request to site you with stream return req: [127.0.0.1:3000/ask/stream?site=you&prompt=who are you]() res: ``` event: message data: {"content":"I"} event: message data: {"content":"'m"} event: message data: {"content":" a"} event: message data: {"content":" search"} event: message data: {"content":" assistant"} ........ event: done data: {"content":"done"} ``` ## 👥 Chat Gr