wechatgpt

introduce

Realize the function of WeChat chat robot and access chatgpt’s latest model gpt-3.5-turbo, which increases the speed by 10 times

Function:

1. Log in to your WeChat account. As a robot, you can reply to private chats or @ messages in the group.
2. It supports automatically sending the login QR code to the designated mailbox. You need to fill in the sender's mailbox key (loginEmail), and the recipient's email address (targetEmail). My code only supports qq mailboxes. You can change the code for other mailboxes. It is very simple. Yes, the nodemailer library supports all mailboxes. Please note that the QR code sent to the mailbox cannot be saved locally. WeChat will fail to recognize the code. You must find another device to display the QR code, and then scan the code with your mobile phone.
3. Support docker deployment
4. support agent
5. Self-defined number of user history conversations

Note: Please use a small account for your WeChat account to prevent accidents. You need to complete your personal information to stay logged in on WeChat. WeChat->Me->Services->Wallet->Identity Information->Personal Information. If there is a complete description after the personal information, it is ok. Yes, this is a necessary step, otherwise it will be easy to disconnect! ! !

Configuration file description

The configuration file supports dynamic loading and can be modified and saved directly to take effect.

 module . exports = {
    email : { // 发送登录二维码到邮箱的配置，掉线时候自动发送登录的二维码到邮箱，随时随地登录
        enable : false , // 为true时 loginEmail targetEmail 必填
        loginEmail : {
            user : '[email protected]' , // 发送的邮箱
            pass : 'xxx' , // 开启IMAP/SMTP服务，提供的密钥，不是邮箱密码
        } ,
        targetEmail : '[email protected]' , // 接收的邮箱
    } ,
    heart : { // 主要是防止微信掉线的，一段时间没有活动，wechaty会自动掉线的 建议开启
        enable : false ,
        timeInterval : 3 * 60 , // 单位秒
        contactName : 'xx' , // 微信定时发送存活消息给指定用户，预防掉线
    } ,
    proxy : {
        enable : false , // 如果使用代理请改为true
        baseURL : 'https://api.openai.com/v1' , //这个是固定的不用修改
        host : '127.0.0.1' , // 修改为自己的代理host
        port : 1080 , // 修改为自己的代理端口
        protocol : 'socks' // 支持http,https,socks
    } ,
    apikey : 'xxxxxxx' , // 你在openai申请的key，如果你没有可以点个star联系我
    limit : {
        all : { // 所有用户的配置，默认值
            historyCount : 3 , // 不同用户，保留历史对话数
            max_tokens : 2048 , // 最大字数限制
        } ,
        user : {
            "contactName" : { // 某个用户的配置，key是用户的微信昵称，不是备注，也不是群昵称，日志里面有打印
                historyCount : 0 , // 针对某些话痨，可以把这个值设置成0，节省api使用字数，设置0即没有历史对话
                max_tokens : 1024 , // 最大字数限制
            } ,
        }
    } ,
}

Startup steps

1. First create a new config.js file in the config directory. You can copy config.demo directly, and then fill in the configuration. The default configuration here is modified and the apikey can be started. Other configurations are optional depending on the situation.

 cp config.demo config.js

Then fill in the configuration

2. start up

Run docker-compose to start the container

 docker-compose up --build -d

You can view the login QR code directly in the log

 docker-compose logs -f wechatgpt

Realize the effect

private chat

Group chat @ it and you can reply