Connect Slack to AI Agents with MCP

The Slack MCP server enables AI agents to interact directly with your Slack workspace. By acting as an automated participant, your agent can read channel histories, post structured updates, manage conversations, and even react to messages with emojis.

Prerequisites

• ▸Slack workspace Administrator or App Manager access.
• ▸A Slack Bot Token (xoxb-...).
• ▸Node.js 18+ installed on your host machine.

Creating a Slack Bot Token

1. ▸Visit api.slack.com/apps.
2. ▸Click Create New App → From scratch.
3. ▸Under OAuth & Permissions, add these Bot Token Scopes:
   • ▸channels:history — Read public channel messages
   • ▸channels:read — List available channels
   • ▸chat:write — Post messages as the bot
   • ▸users:read — View user profiles
   • ▸reactions:write — Add emoji reactions
4. ▸Click Install to Workspace.
5. ▸Copy the generated Bot User OAuth Token.

Configuration Setup

Add the integration to your MCP client config file:

{
  "mcpServers": {
    "slack": {
      "command": "npx",
      "args": ["-y", "@modelcontextprotocol/server-slack"],
      "env": {
        "SLACK_BOT_TOKEN": "xoxb-your-token-here",
        "SLACK_TEAM_ID": "T0123456789"
      }
    }
  }
}

Available Tools

• ▸list_channels — Browse workspace channels.
• ▸post_message — Send messages to channels or individuals.
• ▸reply_to_thread — Reply inside existing message threads.
• ▸add_reaction — React to messages with emojis.
• ▸get_channel_history — Read recent messages from a channel.
• ▸get_thread_replies — Read deep thread conversations.
• ▸search_messages — Full-text search across the workspace.

Example Workflows

• ▸Daily Standup Summary: "Summarize today's messages in the #engineering channel and highlight any action items or blockers."
• ▸Cross-Channel Analysis: "Search for all mentions of 'deployment' across all channels in the last week and create a summary."
• ▸Automated Responses: "Post a message to #general summarizing the key architectural decisions we just made in this chat."

Best Practices & Troubleshooting

• ▸Principle of Least Privilege: Only grant the OAuth scopes your specific use case requires. Do not grant groups:history (private channels) unless absolutely necessary.
• ▸Bot Not In Channel Error: A Slack bot cannot read or post to a channel until it is explicitly invited. Type /invite @YourBotName inside the Slack channel you want it to access.
• ▸Use Threads: Instruct your agent to use the reply_to_thread tool rather than post_message when responding to existing conversations to avoid flooding the channel feed.

Using with OpenAI Codex

You can use this MCP server with the OpenAI Codex CLI by adding it to your configuration:

codex mcp add --name slack --command "npx -y @modelcontextprotocol/server-slack" --env SLACK_BOT_TOKEN=xoxb-your-token

For a full list of recommended servers, see Best MCP Servers for OpenAI Codex.