AI Discord Bot

This repository contains two separate Discord bots, each in its own folder:

Part	Folder	Description
🗨️ General Chat	`general-chat/`	AI chatbot that replies to mentions and DMs using multiple AI models with automatic fallback. Responses stream to Discord in real-time.
🏢 AI Company	`ai-company/`	Simulates a company discussion: give it a task, pick roles (CEO, CTO, Designer, …), and receive a structured Final Outcome. See `ai-company/SETUP.md`.

Both bots are powered by Pollinations AI and run on Python 3.11.
Each has its own GitHub Actions workflow so they can be started independently.

For platform-specific deployment instructions (Linux, macOS, Windows, Docker, GitHub Actions) see PLATFORM.md.
For Vercel and always-on hosting options see VERCEL.md.
For Railway.com (easiest always-on hosting) see RAILWAY.md.

Repository Structure

AI-discord-bot/
├── general-chat/
│   ├── bot.py              # General-purpose AI chat bot (streaming responses)
│   └── requirements.txt
├── ai-company/
│   ├── bot.py              # AI company simulation bot (interactive mode)
│   ├── requirements.txt
│   └── SETUP.md            # Detailed setup guide for the AI company bot
├── .github/workflows/
│   ├── discord-bot.yml             # Workflow: General Chat bot
│   └── discord-ai-company-bot.yml  # Workflow: AI Company bot
├── PLATFORM.md             # Platform-specific deployment guide
├── VERCEL.md               # Vercel / always-on hosting guide
├── RAILWAY.md              # Railway.com deployment guide (easiest always-on)
└── README.md

Part 1 — General Chat Bot (`general-chat/`)

An AI-powered Discord bot that responds to mentions and direct messages using multiple AI models from Pollinations AI. It automatically tries a chain of models (OpenAI, Gemini, Claude, DeepSeek, and more) so you always get a reply even when one model is unavailable.

Responses are streamed to Discord in real-time — the bot sends a placeholder message immediately and progressively updates it as tokens arrive, so you see the answer forming word-by-word instead of waiting for the entire reply.

GitHub Actions secrets required: DISCORD_TOKEN, POLLINATIONS_TOKEN

Prerequisites

Requirement	Notes
Python 3.11+	Only needed for local deployment
A Discord account	To create a bot application
A Pollinations account	To obtain an API key

Step 1 – Create a Discord Application & Bot

Go to the Discord Developer Portal and click New Application.
Give your application a name (e.g. My AI Bot) and click Create.
In the left sidebar click Bot, then click Add Bot → Yes, do it!
Under the TOKEN section click Reset Token, then Copy the token and save it somewhere safe – this is your DISCORD_TOKEN.
Scroll down to Privileged Gateway Intents and enable Message Content Intent (the bot needs this to read messages). Click Save Changes.

Step 2 – Get a Pollinations API Key

Go to https://enter.pollinations.ai and sign in.
Create or copy your API key – this is your POLLINATIONS_TOKEN.

Step 3 – Invite the Bot to Your Server

In the Discord Developer Portal, open your application and click OAuth2 → URL Generator in the left sidebar.
Under Scopes check bot and applications.commands (required for the /ask slash command).
Under Bot Permissions check at minimum:
- Read Messages / View Channels
- Send Messages
- Read Message History
Copy the generated URL, open it in your browser, and select the server you want to add the bot to.

Deployment

Option A – Run Locally

# 1. Clone the repository
git clone https://github.com/hugow0528/AI-discord-bot.git
cd AI-discord-bot

# 2. Install dependencies
pip install -r general-chat/requirements.txt

# 3. Set environment variables
export DISCORD_TOKEN="your_discord_bot_token_here"
export POLLINATIONS_TOKEN="your_pollinations_api_key_here"

# Windows (Command Prompt)
# set DISCORD_TOKEN=your_discord_bot_token_here
# set POLLINATIONS_TOKEN=your_pollinations_api_key_here

# 4. Start the bot
python general-chat/bot.py

You should see a log line like:

2025-01-01 00:00:00,000 [INFO] __main__: Logged in as MyAIBot#1234 (ID: 123456789012345678)

The bot is now online. Keep the terminal window running; closing it will stop the bot.

Option B – GitHub Actions

GitHub Actions can host the bot with zero infrastructure setup. The included workflow (.github/workflows/discord-bot.yml) lets you start the bot manually via the Actions tab.

Note: The automatic 6-hour cron restart has been disabled. GitHub Actions jobs have a maximum runtime of ~6 hours. For continuous uptime use Railway, Render, or a VPS (see below).

Fork or push this repository to your own GitHub account.
Add Secrets – In your repository go to Settings → Secrets and variables → Actions → New repository secret and add:

Name Value

DISCORD_TOKEN Your Discord bot token from Step 1

POLLINATIONS_TOKEN Your Pollinations API key from Step 2
Enable the workflow – Go to the Actions tab in your repository. If prompted, click I understand my workflows, go ahead and enable them.
Run the bot – Click on Discord AI Bot in the workflows list, then click Run workflow → Run workflow. The bot will come online within a minute.

Tip: You can monitor the bot's logs in real-time from the Actions tab by clicking the running workflow job.

Option C – Always-On Hosting (Vercel / Railway / Render)

For a bot that never goes offline:

RAILWAY.md — Railway (recommended, easiest always-on option; full gateway bot, free starter credit)
VERCEL.md — Vercel (HTTP Interactions / slash commands only), Render (full gateway bot, free tier available)

How to Use the Bot

Once the bot is online and in your server, you can interact with it in several ways.

1. Basic Usage

Mention the bot in any channel it can read, followed by your question:

@MyAIBot What is the capital of France?

The bot automatically selects the best available model and replies directly to your message.

2. Get Help (`@about`)

Mention the bot with @about to receive a full usage guide and a list of all available @model prefixes:

@MyAIBot @about

The bot will reply with a formatted message describing all features, available models, and the /ask slash command.

3. Choose a Specific Model

Prefix your message with @<model-name> (right after mentioning the bot) to tell it which AI model to use. If that model fails, the bot automatically falls back through the rest of the chain.

@MyAIBot @deepseek Explain the theory of relativity.
@MyAIBot @gemini-search What's in the news today?
@MyAIBot @claude-fast Write a haiku about autumn.

You can also use @ai as an explicit alias for "use any available model" (the default behaviour):

@MyAIBot @ai What is 2 + 2?

4. Slash Commands

Type / in any channel to see the bot's available slash commands. Discord shows them above the text input as you type.

Command	Parameters	Description
`/ask`	`question` (required), `model` (optional)	Ask the AI a question; response streams to Discord in real-time.
`/interrupt`	`perspective` (required)	Inject your point of view mid-request and have the AI continue from your perspective instead of hard-cancelling.
`/cancel`	—	Hard-cancel your current in-progress AI request.
`/about`	—	Show the full usage guide and model list.
`/models`	—	List all available AI models with vision support info.
`/settings`	`model` (optional)	View or set your preferred AI model.

When you use /ask, click the model field to instantly see the full list of available models as a dropdown — no typing required. Select a model and press Enter to confirm, or leave it blank for automatic selection.

/ask question:What is quantum computing?
/ask question:What is in the news today? model:gemini-search

Streaming responses

Responses from /ask and @mention messages are streamed to Discord in real-time. As soon as the AI starts generating, the bot sends a placeholder message (▌) and updates it progressively — you see the reply form word-by-word instead of waiting for the complete response.

Injecting your perspective (`/interrupt`)

While the AI is generating a response (or at any time), use /interrupt to inject your own point of view and redirect the conversation:

/interrupt perspective:Focus more on security implications
/interrupt perspective:Actually, I meant for a mobile app, not web

The bot cancels any running request, adds your perspective to the conversation context, and generates a new response that takes your input into account.

3. Image / Vision Input

Attach one or more images to your message and the bot will analyse them using a vision-capable model. You can combine text and images:

@MyAIBot @openai-fast What breed is this dog?   ← with an image attached
@MyAIBot Describe this chart.                   ← with an image attached

If you attach an image without any text, the bot defaults to "Describe this image."

Vision-capable models: openai, openai-fast, gemini, gemini-fast, gemini-large, gemini-search, claude-fast

If you explicitly pick a non-vision model (e.g. @deepseek) and attach an image, the image is silently ignored and only the text is sent.

4. Direct Messages (DM)

Open a DM with the bot and send any message – no mention required. The same @<model-name> prefix and image-attachment features work here too:

@deepseek Summarise this article for me.

Available Models

Model name	Supports images?	Notes
`openai-fast`	✅	Fast OpenAI model; default first choice
`openai`	✅	Full OpenAI model
`gemini-search`	✅	Gemini with Google Search grounding
`gemini`	✅	Standard Gemini model
`gemini-fast`	✅	Faster Gemini variant
`gemini-large`	✅	Larger Gemini variant
`claude-fast`	✅	Fast Claude model
`glm`	❌	GLM model
`qwen-character`	❌	Qwen character model
`deepseek`	❌	DeepSeek model
`qwen-safety`	❌	Qwen safety model

The default fallback order (when no model is specified) is: openai-fast → gemini-search → openai-fast → openai → glm → claude-fast → qwen-character → deepseek → qwen-safety

Example Interactions

You type	What happens
`@MyAIBot Explain quantum computing`	Uses the default model chain
`@MyAIBot @ai Explain quantum computing`	Same as above (`@ai` = default chain)
`@MyAIBot @about`	Returns the full usage guide and model list
`@MyAIBot @gemini-search What's in the news?`	Uses `gemini-search` first, then falls back if needed
`@MyAIBot @openai-fast` + image attached	Analyses the image with `openai-fast`
`@MyAIBot` + image (no text)	Sends "Describe this image." to the model
(In DM) `@deepseek Write a poem`	Uses `deepseek` first in DM
`/ask question:Hello model:deepseek`	Slash command; model picked from dropdown
`/about`	Shows the full usage guide via slash command
`/models`	Lists all AI models via slash command

Other Notes

The bot ignores its own messages to prevent feedback loops.
Responses are streamed in real-time — the bot edits a placeholder message progressively as tokens arrive.
If a response exceeds Discord's 2,000-character limit, the bot automatically splits it into multiple messages.
If all models fail, the bot replies: "Sorry, all AI models are currently unavailable. Please try again later."

Configuration Reference

All configuration is done through environment variables:

Variable	Required	Description
`DISCORD_TOKEN`	✅	Discord bot token obtained from the Developer Portal
`POLLINATIONS_TOKEN`	✅	Pollinations AI API key from https://enter.pollinations.ai

The AI API endpoint (https://gen.pollinations.ai/v1/chat/completions) and the model chain are defined at the top of general-chat/bot.py and can be customised there if needed.

Part 2 — AI Company Bot (`ai-company/`)

For full setup and usage instructions, see ai-company/SETUP.md.

Quick Overview

Give the bot a task and it will simulate a company brainstorming session across multiple roles, then produce a structured Final Outcome.

GitHub Actions secrets required: DISCORD_TOKEN_COMPANY, POLLINATIONS_TOKEN

Slash Commands

Command	Parameters	Description
`/company`	`task` (required), `roles` (optional), `interactive` (optional)	Run a multi-role AI company discussion
`/build`	`task` (required), `project_name` (optional), `roles` (optional), `interactive` (optional)	Developer team discussion + code generation saved to `project/`
`/followup`	`request` (required)	Ask questions or request amendments after a `/build` session
`/company_roles`	—	List all available and custom roles
`/about`	—	Show the help guide

Interactive Mode — Being an Interruptor

Pass interactive:True to /company or /build to become a participant in the discussion. After each role responds, the bot shows two buttons:

▶ Continue — let the discussion proceed.
✏️ Add My Input — open a text modal to type your perspective (up to 1 000 characters). Your input becomes Stakeholder Input that all subsequent roles and the Facilitator can see and react to, steering the discussion from your point of view.

Example

/company task:Build a food delivery mobile app

The bot will produce something like:

🏢 AI Company Discussion
📋 Task: Build a food delivery mobile app
👥 Participants: CEO, CTO, Product Manager, Designer, Engineer, Marketing Manager

👤 CEO
We should focus on the top-3 metro areas first to validate PMF before scaling nationally...

👤 CTO
A React Native app backed by a Node.js microservice architecture would let us ship cross-platform quickly...

... (each role contributes) ...

---
✅ Final Outcome

Key decisions:
1. Target three cities for the MVP launch
2. Use React Native + Node.js microservices

Next steps:
1. Define MVP feature set with the Product Manager
2. Hire a mobile engineer and UX researcher
...

Custom Roles

You can supply any role name — even ones not built into the bot:

/company task:Design a new city transport system roles:Urban Planner,Civil Engineer,Mayor,Environmentalist

Run the AI Company Bot Locally

pip install -r ai-company/requirements.txt
export DISCORD_TOKEN_COMPANY="your_company_bot_token"
export POLLINATIONS_TOKEN="your_pollinations_api_key"
python ai-company/bot.py

Name		Name	Last commit message	Last commit date
Latest commit History 7 Commits
Project		Project
ai-company		ai-company
general-chat		general-chat
APIDOCS.md		APIDOCS.md
PLATFORM.md		PLATFORM.md
RAILWAY.md		RAILWAY.md
README.md		README.md
VERCEL.md		VERCEL.md

Name	Value
`DISCORD_TOKEN`	Your Discord bot token from Step 1
`POLLINATIONS_TOKEN`	Your Pollinations API key from Step 2

Folders and files

Latest commit

History

Repository files navigation

AI Discord Bot

Repository Structure

Part 1 — General Chat Bot (general-chat/)

Table of Contents

Prerequisites

Step 1 – Create a Discord Application & Bot

Step 2 – Get a Pollinations API Key

Step 3 – Invite the Bot to Your Server

Deployment

Option A – Run Locally

Option B – GitHub Actions

Option C – Always-On Hosting (Vercel / Railway / Render)

How to Use the Bot

1. Basic Usage

2. Get Help (@about)

3. Choose a Specific Model

4. Slash Commands

Streaming responses

Injecting your perspective (/interrupt)

3. Image / Vision Input

4. Direct Messages (DM)

Available Models

Example Interactions

Other Notes

Configuration Reference

Part 2 — AI Company Bot (ai-company/)

Quick Overview

Slash Commands

Interactive Mode — Being an Interruptor

Example

Custom Roles

Run the AI Company Bot Locally

About

Resources

Uh oh!

Stars

Watchers

Forks

Releases

Packages 0

Uh oh!

Contributors

Uh oh!

Languages

Part 1 — General Chat Bot (`general-chat/`)

2. Get Help (`@about`)

Injecting your perspective (`/interrupt`)

Part 2 — AI Company Bot (`ai-company/`)

Packages