chatgpt-html

Introduction to ChatGPT-HTML

ChatGPT-HTML is an innovative web-based project designed for integrating ChatGPT's official interface into a simple HTML online chat platform. It was developed using a proxy of the ChatGPT web version, providing a free alternative for those who wish not to use the paid ChatGPT API KEY. However, if one is interested in using the API KEY, a visit to chatgpt-web is recommended.

Important Considerations

A critical prerequisite for utilizing this project is ensuring your server's IP has not been banned. Unfortunately, this was the case with the developer's server, making it impossible for them to verify the project's operability. So, users are encouraged to try it on their own.

Key Features

• The project features a straightforward file structure, making it highly accessible for beginners.
• It provides essential functionalities such as continuous conversations, multi-user session isolation, and markdown format output.

Deployment Process

Using Docker Compose

To get started with deployment, all necessary files should be placed in the same directory. Here's a step-by-step guide:

1. Create a config.json File: This file is fundamental for setting up the configuration. It should look like this:

   {
       "__comment01__": "Choose one among email, session_token, and access_token",
       "__comment02__": "Email authentication (only supports accounts registered the standard way, not through Google or Microsoft)",
       "email": "",
       "password": "",
       "__comment03__": "Session token authentication (unaffected by registration method)",
       "session_token": "",
       "__comment04__": "Access token authentication (unaffected by registration method)",
       "access_token": "",
       
       "__comment05__": "The following are optional fields",
       "__comment06__": "Connect to the proxy endpoint (the author's server was blocked, proxies seem to only work with no-password SOCKS5 or HTTP; please test yourself)",
       "proxy": "",
       "__comment07__": "Use a paid OpenAI account (advertised as faster, with no frequency limit, change false to true)",
       "paid": false
   }
   

2. Obtain Session and Access Tokens: If using tokens for authentication, follow these methods:

   • Session Token: Visit https://chat.openai.com/chat, open developer tools (F12), and find the __Secure-next-auth.session-token cookie. Copy its value.
   • Access Token: Log into the ChatGPT official web version, then open https://chat.openai.com/api/auth/session to obtain the token.

3. Create docker-compose.yml Configuration: Set up Docker Compose with the following contents:

   services:
     chatgpt:
       image: sheepgreen/chatgpt-html
       container_name: htmchat
   #    environment:
   #      - CHATGPT_BASE_URL=your proxy server address (if left empty, it defaults to the author's server, which may be intermittently unavailable)
       volumes:
         - ./config.json:/chatgpt-html/config.json
   #      - ./chat.html:/chatgpt-html/templates/chat.html # By default, my UI is built-in. Uncomment to replace it with your own webpage
       ports:
         - "9999:8088" # Port 8088 is internal and should not be changed; port 9999 is external and adjustable
       restart: always
   

4. Start the Project: Run docker-compose up -d to launch the application successfully.

Important Notes

• Access the chat through http://ip:port.
• After modifying the chat.html file, execute docker restart htmchat for changes to take effect.

Additional Resources

The project is part of a broader ecosystem of related tools, including:

• ChatGPT Telegram Bot: link
• ChatGPT Enterprise WeChat Application Bot: link
• QQ Channel ChatGPT Docker Version: link
• Microsoft Bing Telegram Bot Docker Version: link
• Google BARD Web Version: link
• Google BARD Telegram Bot: link

For anyone intrigued by experimenting with Bing's chat feature, an experimental test site was created using waylaidwanderer's approach at test site. This site allows model switching by clicking the icon beside the chat box, offering models like Bing (GPT-4) and Sydney, which has fewer interaction limitations.

By leveraging the pandora project, an OpenAI ChatGPT high imitation site can also be accessed here.

This introduction serves as an overview for users keen on exploring the extensive capabilities of ChatGPT-HTML, offering a seamless entry into the world of AI-powered chat functionalities.