How to use the Slack MCP server with Claude flawlessly

I spend most of my day in Slack. Team updates, meeting notes, quick decisions, and follow-ups. It all lives there. Threads move fast, and channels fill up even faster. When I want to use that context with an AI, I end up copying pieces into a prompt, refining them, and explaining what the thread was about.

I have done that more times than I can count. It always feels clunky.

Slack MCP changes that. It connects directly to your workspace and lets tools work with real conversations in real time using the actual content from your channels and threads.

Once connected, you can use the Slack data you already have and respond with the context that matters.

In this post, I will show you how to set up Slack MCP and use it to make Claude more helpful with your day-to-day conversations.

What is MCP?

Think of MCP as a bridge that connects all your SaaS tools to your AI agent. It acts like an adapter, enabling your AI agent (Client) to understand and interact with your tools.

According to Anthropic (the team behind Claude and MCP),

“MCP is an open protocol that standardizes how applications provide context to LLMs. Think of MCP like a USB-C port for AI applications. Just as USB-C provides a standardized way to connect your devices to various peripherals and accessories, MCP provides a standardized way to connect AI models to different data sources and tools.”

credits: modelcontextprotocol.io

For more, check out this detailed post or MCP components: MCP Esplained

What is Slack MCP

Slack MCP is a secure integration that connects your Slack workspace to MCP clients, such as Claude and Cursor. It gives large language models direct access to your real Slack data through a fixed set of supported actions.

Once connected, Slack MCP allows these tools to:

• Read messages from channels or threads: Pull recent messages from public or private Slack channels for quick reference or summaries.
• Post messages: Send new messages into any Slack channel or thread based on prompts or actions.
• Reply in threads: Add follow-up responses to existing conversations in a thread format.
• Search messages: Find messages by keyword, timestamp, or sender, directly through the model.
• List users or channels: Retrieve a list of members or available channels in your Slack workspace.
• Fetch user info: Get details about any team member to personalize messages or automate updates.
• and more ….

This setup turns Slack into a live data source. Instead of copying and explaining messages to your tools, you can let them interact with your workspace directly, based on actual, up-to-date information.

Setting Up Slack MCP with Claude

To connect Claude with your Slack workspace, you’ll first need to set up Slack MCP using Composio. This allows Claude to access Slack messages, threads, and channels through a secure and permissioned connection.

Once the connection is active, you can use Slack content directly in your prompts without copying messages or jumping between tools.

Prerequisites

Before you begin, make sure the following are ready on your system:

• Claude Desktop

  Download it from https://claude.ai/download and install the right version for your OS.

• Node.js

  You’ll need Node installed to run the setup script. Run node -v in your terminal to confirm. If it’s missing, install it from https://nodejs.org.

Step 1: Get the Setup Script from Composio

Head to https://mcp.composio.dev

Search for Slack and open the integration page

Look for the setup script under the Claude tab, It should look something like this:

npx @composio/mcp@latest setup "https://mcp.composio.dev/partner/composio/slack/mcp?customerId=<your_customer_id>" "" --client claude 

Copy this script, you will need it for the next step

Step 2: Check the Available Actions

Once you're on the Slack MCP setup page in Composio, make sure these 10 supported actions are enabled. These define what your assistant will be able to do inside Slack:

• Add reaction to message
• Create a reminder
• Fetch conversation history
• List conversations
• List all Slack team users with pagination
• Remove reaction from item
• Schedule message in chat
• Search Messages
• Send a message to a Slack channel
• Update a Slack message

These actions give real-time access to messages, channels, users, and scheduling. You only need to enable the ones your workflow will actually use.

Step 3: Connect Slack from the Claude Config Folder

Launch Claude Desktop

Go to File > Settings, then open Developer > Edit Config

This will open the config directory on your system. On Windows, the path is usually:

``` C:\Users\<your_username>\AppData\Roaming\Claude ``` 

In that folder, open a terminal window and paste and run the script you copied earlier

``` npx @composio/mcp@latest setup "https://mcp.composio.dev/partner/composio/slack/mcp?customerId=<your_customer_id>" "" --client claude ``` 

This will launch an OAuth2 flow in the terminal. You will be redirected to sign into your slack account and accept permissions.

Once authentication is complete, you can see a screen like this

In Terminal, you can find this message

Step 4: Confirm the Connection (Optional)

In the same config folder, open:

claude_desktop_config.json 

You should see Slack listed under "clients" With the connected MCP endpoint. That confirms the setup worked.

Step 5: Restart Claude

After completing the setup, close and reopen Claude Desktop to activate the Slack MCP integration.

Once Claude restarts, open the panel (as shown in the image) to confirm the connection.

You will see your Slack MCP integration listed with the name you gave it during setup. For example, if you named it Morning agent, that name will appear in Claude.

Claude is now ready to use your Slack workspace data in real time, so you can reference channels, threads, and messages directly in your prompts.

Use Case: Summarize Long Messages and Follow Up in Slack Threads

In busy Slack channels, long-form messages and updates often contain important information, decisions, announcements, event details, or kudos. With Slack MCP connected, you can use Claude to automatically summarize those messages and post follow-ups directly in the same threads.

Prompt Example

1. Read the most recent long messages or threads in #general and #social. 2. Summarize key points: announcements, actions, and acknowledgments. 3. Reply to each thread with a summary and any next steps. 

🔧 What Happens Behind the Scenes (via Slack MCP)

Why Use This?

• Reduce cognitive load: Get quick insights without reading every message
• Stay aligned: Team members can catch up even if they missed the original post
• Move faster: Follow-ups are generated and posted automatically

Watch it in action

This video shows the full flow from reading real Slack messages to summarizing them and posting a follow-up directly in the thread using Slack MCP.

Summary

Slack MCP lets Claude access and interact with your Slack workspace in real time. Instead of copying and pasting threads into prompts, you can let Claude read messages, summarize conversations, post replies, and keep you updated directly in your Slack.

By following a few simple steps using Composio, you can securely connect Slack to Claude. Once set up, Claude becomes a helpful assistant, working right alongside you in your Slack channels.

Frequently Asked Questions

1. Where can I manage or disconnect the Slack connection I created with Composio?

You can manage your active connections at https://mcp.composio.dev. After signing in, go to your Slack integration, where you'll find options to revoke, rename, or reconfigure your existing setup.

2. Can I connect multiple Slack workspaces using Composio?

Yes, Composio allows you to set up multiple Slack MCP instances. Each workspace will appear as a separate client, and you can name them individually (e.g., “Morning agent” or “Product HQ”) to keep things organized in Claude.

3. Will Claude post in Slack without my permission?

No. Claude only takes actions based on your prompts. It does not automatically send or post messages in Slack unless you specifically ask it to.

4. Can Claude summarize private channels or DMs using Slack MCP?

Yes, if the Slack app (via Composio) was authorized with access to private channels or DMs, Claude can read those conversations as part of the integration. Only channels to which you have access will be visible to the model.

5. What happens if a message or thread gets edited after Claude has summarized it?

Claude will only reflect the content available at the time it fetched the messages. If a conversation evolves, you can simply run the prompt again, and Claude will pull the latest version of the thread using the Slack MCP connection.