ChatGPT Discord Bot

Build your own Discord bot using ChatGPT

[!IMPORTANT]

Major Update (2024/03):

• Due to instability issues with GPT-4 model, we have defaulted back to GPT-3.5-turbo
• Gemini-Pro and GPT-4 now supported for free
• See README for details and dependency updates.

Chat

Setup

Prerequisites

• Python 3.9 or later
• Rename the file .env.example to .env
• Running pip3 install -r requirements.txt to install the required dependencies
• Google Chrome for Image Generation

Step 1: Create a Discord bot

1. Go to https://discord.com/developers/applications create an application

2. Build a Discord bot under the application

3. Get the token from bot setting

   

4. Store the token to .env under the DISCORD_BOT_TOKEN

   

5. Turn MESSAGE CONTENT INTENT ON

   

6. Invite your bot to your server via OAuth2 URL Generator

   

Optional: OPENAI ACCESS (gpt-4 & DALL-E-3 supported)

If you possess an OpenAI Plus account, utilizing the OpenAI provider is recommended for a smoother experience.

1. Navigate to https://chat.openai.com/ and log in with your account credentials.

2. Launch the Developer Tools in your browser.

3. In the Developer Tools, navigate to the Network tab.

4. Refresh the page to record the network activity.

5. Chat whth the bot one time.

6. In the Network tab, right-click on any item in the list of network activities and choose Save all as HAR with content. Save the file with the name a.har.

7. place the a.har file in the ./hardir directory.

• You can change MODEL in .env to gpt-4 also

[!CAUTION] Ensure that your a.har file is stored securely, as it may contain sensitive information.

This is not OpenAI API. For API access, refer Optional: Configuring OpenAI API

Step 2: Run the bot on the desktop

1. Open a terminal or command prompt

2. Navigate to the directory where you installed the ChatGPT Discord bot

3. Run python3 main.py or python main.py to run the bot

Step 2: Run the bot with Docker

1. Build the Docker image & run the Docker container with docker compose up -d

2. Inspect whether the bot works well docker logs -t chatgpt-discord-bot

   Stop the bot:

   • docker ps to see the list of running services
   • docker stop <BOT CONTAINER ID> to stop the running bot

Have a good chat!

Image Generation

OpenAI DALLE3 Image Generation (Requires a GPT Plus account)

1. Log into your openai account

2. Go to https://chat.openai.com/api/auth/session

3. Copy the value for access_token and paste it into .env under OPENAI_TOKEN

Microsoft Bing Image Generation

1. Go to https://www.bing.com/chat and log in

2. Open console with F12

3. Open Application tab > Cookies

4. Copy the value for _U from cookies and paste it into .env under BING_COOKIE

Google Gemini Image Generation

1. Go to https://gemini.google.com/app and log in

2. Open console with F12

3. Open Application tab > Cookies

4. Copy the value for __Secure-1PSID from cookies and paste it into .env under GOOGLE_PSID

Optional: Configuring OpenAI API

To use the OpenAI API features, follow these steps:

1. Obtain your API key by visiting https://platform.openai.com/api-keys
2. Paste the API key under OPENAI_KEY in .env
3. Set OPENAI_ENABLED to True in .env

[!NOTE] GPT-4 API is subject to certain restrictions.

For more details, please visit https://help.openai.com/en/articles/7102672-how-can-i-access-gpt-4

Optional: Setup system prompt

• A system prompt would be invoked when the bot is first started or reset

• You can set it up by modifying the content in system_prompt.txt

• All the text in the file will be fired as a prompt to the bot

• Get the first message from ChatGPT in your discord channel!

• Go Discord setting turn developer mode on

  1. Right-click the channel you want to recieve the message, Copy ID

     

  2. paste it into .env under DISCORD_CHANNEL_ID

Optional: Disable logging

• Set the value of LOGGING in the .env to False

中文設置教學

Commands

• /chat [message] Chat with ChatGPT/Gemeni

• /draw [prompt] Generate an image with Gemini/OpenAI/Bing

• /switchpersona [persona] Switch between optional chatGPT jailbreaks

  • random: Picks a random persona
  • standard: Standard chatGPT mode
  • dan: DAN 13.5 (Latest Working ChatGPT Jailbreak prompt)
  • Smart mode: AIM (Always Intelligent and Machiavellian)
  • Developer Mode: software developer who specializes in the AI's area

• /private ChatGPT switch to private mode

• /public ChatGPT switch to public mode

• /replyall ChatGPT switch between replyAll mode and default mode

• /reset Clear ChatGPT conversation history

• /chat-model Switch different chat model

  • gpt-4: GPT-4 model
  • Gemini: Google Gemini Model

Special Features

Switch Persona

Warning

Certain personas may generate vulgar or disturbing content. Use at your own risk.

Mode

• public mode (default) the bot directly reply on the channel

  

• private mode the bot's reply can only be seen by the person who used the command

  

• replyall mode the bot will reply to all messages in the channel without using slash commands (/chat will also be unavailable)

  Warning The bot will easily be triggered in replyall mode, which could cause program failures