MCP Discord Agent Communication
An MCP server that enables AI agents to send notifications and request user input via Discord during long-running tasks. It allows users to remotely interact with their AI assistants and provide feedback through the Discord messaging platform.
README
MCP Discord Agent Communication
An MCP (Model Context Protocol) server that enables async communication between AI agents and users via Discord. Perfect for long-running AI tasks where you need remote interaction capabilities.
🎯 Purpose
This isn't a full Discord bot - it's specifically designed for AI agents to communicate with users remotely during long-running tasks. Step away from your machine while your AI assistant handles complex work and reaches out when needed.
Features
- 💬 Unified Communication - Single tool for all Discord interactions
- 📢 Send Notifications - Agent sends updates without waiting for response
- 🔄 Request User Input - Agent asks questions and waits for replies
- ⏰ Async Communication - Perfect for long-running tasks requiring remote interaction
Use Cases
Perfect for scenarios where AI agents need remote human interaction:
- Kick off complex tasks and get notified when they complete
- Let AI ask for your input or feedback when decisions are needed
- Stay in the loop while doing other activities (workouts, walks, meetings)
- Monitor progress without being glued to your screen
Quick Setup
1. Create Discord Bot
- Go to Discord Developer Portal
- Create "New Application" → Go to "Bot" → "Add Bot"
- Copy the bot token
- Enable "MESSAGE CONTENT INTENT" in bot settings
- Use OAuth2 URL Generator to invite bot to your server with these permissions:
- Send Messages, Read Message History, Add Reactions, View Channels
2. Get Channel ID
Enable Developer Mode in Discord settings, then right-click your channel → "Copy Channel ID"
3. Configure Claude Desktop
Add to ~/Library/Application Support/Claude/claude_desktop_config.json (macOS) or %APPDATA%\Claude\claude_desktop_config.json (Windows):
{
"mcpServers": {
"discord-agent-comm": {
"command": "npx",
"args": ["mcp-discord-agent-comm"],
"env": {
"DISCORD_BOT_TOKEN": "your-bot-token-here",
"DISCORD_CHANNEL_ID": "your-channel-id-here"
}
}
}
}
Available Tool
discord_message
Send a message to Discord, optionally waiting for a response.
Simple notification (no response needed):
{
"message": "Task completed! 🎉"
}
Request user input (waits for reply):
{
"message": "What should I name this file?",
"expect_reply": true
}
Yes/No or any other question (waits for reply):
{
"message": "Should I deploy to production?",
"expect_reply": true
}
Note: When expect_reply is true, the tool waits for the user to reply to the message in Discord.
Agent Instructions
For guidance on how to instruct AI agents to use these tools effectively, see AGENT_INSTRUCTIONS.md. It provides example guidelines for when agents should communicate vs. work autonomously.
Important Notes
Discord Notifications
- Only one device receives Discord notifications - disable Discord on other devices or turn off notifications if you want them on a specific device
- Notifications work best on mobile devices when other instances are closed
How It Works
- Notifications: Direct message to channel (when
expect_replyis false or omitted) - Input requests: Users reply to the bot's message using Discord's reply feature (when
expect_replyis true) - Default timeout: 300 seconds (5 minutes) when waiting for replies
Troubleshooting
- Bot offline: Check token and permissions
- No notifications: Ensure MESSAGE CONTENT INTENT enabled, check channel permissions
- Timeouts: Users must reply to messages (not just type in channel) or click reactions
Development
git clone https://github.com/EugenEistrach/mcp-discord-agent-comm.git
cd mcp-discord-agent-comm
bun install
cp .env.example .env # Edit with your tokens
bun run dev
License
MIT
Publishing Status
Install globally: npm install -g mcp-discord-agent-comm
推荐服务器
Baidu Map
百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。
Playwright MCP Server
一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。
Magic Component Platform (MCP)
一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。
Audiense Insights MCP Server
通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。
VeyraX
一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。
graphlit-mcp-server
模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。
Kagi MCP Server
一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。
e2b-mcp-server
使用 MCP 通过 e2b 运行代码。
Neon MCP Server
用于与 Neon 管理 API 和数据库交互的 MCP 服务器
Exa MCP Server
模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。