MCP Servers

Slack MCP Server — Connect AI to Your Workspace

Set up and use the Slack MCP server to let Claude search messages, read channels, post updates, and summarize conversations in your Slack workspace.

What the Slack MCP Server Does

The Slack MCP server gives AI assistants direct access to your Slack workspace through the Model Context Protocol. Instead of copying and pasting messages into a chat window, you can ask Claude to search your Slack history, read specific channels, summarize long threads, find messages from particular people, and draft or send responses — all through natural language.

This is especially powerful for teams that use Slack as their primary communication hub. Information trapped in Slack threads — decisions, context, technical discussions, customer feedback — becomes instantly accessible to the AI. You can ask questions like "What did the engineering team decide about the migration last week?" and get an answer pulled directly from the relevant channels.

Installation and Configuration

The Slack MCP server requires a Slack app with appropriate API permissions. You will need to create a Slack app in your workspace's admin panel, configure the required OAuth scopes, and install it to your workspace.

First, create a new Slack app at api.slack.com/apps. Under OAuth & Permissions, add these bot token scopes:

channels:history    — Read messages in public channels
channels:read      — List public channels and their metadata
groups:history      — Read messages in private channels the bot is in
groups:read         — List private channels the bot is in
search:read         — Search messages across the workspace
users:read          — Look up user profiles and names
chat:write          — Send messages (optional, for posting)
reactions:read      — Read emoji reactions on messages

Install the app to your workspace and copy the Bot User OAuth Token (starts with xoxb-). This token authenticates the MCP server with Slack's API.

Connecting to Claude Desktop

Add the Slack MCP server to your Claude Desktop configuration. The official Slack MCP server is available as an npm package:

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

Restart Claude Desktop after saving the configuration. The Slack tools should appear in the MCP tool panel. You can verify the connection by asking Claude to list your Slack channels.

Available Tools

The Slack MCP server typically exposes these tools for the AI to use:

search_messages searches across all accessible channels using Slack's search API. This supports the same query syntax as Slack's search bar — you can filter by channel, user, date range, and keywords. The AI uses this when you ask open-ended questions about what happened or what was discussed.

read_channel retrieves recent messages from a specific channel. This is useful for catching up on a channel, reviewing recent activity, or getting context before a meeting. The tool returns messages with timestamps, authors, thread indicators, and reactions.

read_thread fetches all replies in a specific thread. Slack threads often contain the most valuable discussion — decisions, technical details, action items — and this tool makes them accessible without manually navigating to each one.

send_message posts a message to a channel or thread on your behalf. The AI will always confirm before sending, showing you the message content and destination. This is useful for drafting standup updates, responding to questions, or posting automated summaries.

Practical Use Cases

The most valuable use case is information retrieval across a large workspace. In organizations with hundreds of channels, finding specific decisions, technical discussions, or customer mentions is time-consuming. With the Slack MCP server, you can ask Claude to find relevant discussions across all channels simultaneously — something that would take significant manual searching.

Meeting preparation is another strong use case. Before a project sync, ask Claude to summarize all recent activity in the project's Slack channel, identify open questions, and flag any blockers mentioned in threads. You walk into the meeting with full context instead of spending the first ten minutes catching up.

For AI detection and content authenticity work, the Slack MCP server enables workflow automation. An analyst can ask Claude to search for all mentions of a specific article or URL across the workspace, pull the discussion context, and cross-reference it with detection tool results — all without leaving the conversation.

Security Considerations

The Slack MCP server has access to everything the bot token can reach. Review the OAuth scopes carefully and only grant the minimum permissions needed. If you only need read access, omit the chat:write scope. Store the bot token securely — treat it like any other API credential. Do not commit it to version control or share it in plain text.

For enterprise deployments, consider running the MCP server behind your organization's identity provider with additional authentication. The MCP specification supports OAuth-based auth for remote servers, which lets you enforce your organization's access policies on top of the Slack token permissions. Monitor usage through Slack's audit logs to track which channels and messages the bot accesses.

Last updated: 2026 • Browse all courses