Slack

Slack is an MCP server that integrates Slack workspaces with Model Context Protocol clients for searching, reading, and sending workspace communications. Built for developers, operations teams, and workplace analysts, this integration connects AI assistants directly to public channels, private channels, direct messages, and group DMs. It allows assistants to retrieve conversation histories by date ranges or message counts, trace specific discussion threads, list channels, and search messages using precise user and timeframe filters. The server operates over Stdio, SSE, and HTTP transports and provides support for Enterprise setups and proxy configurations. Security-conscious deployments can run in an OAuth mode or a stealth mode that requires no workspace bot permissions or administrative approvals. For safety, tools that modify state, such as posting messages or adding and removing emoji reactions, are disabled by default and must be explicitly enabled via environment variables with optional per-channel restrictions.

Category: Communication & Messaging

Tags: collaboration, messaging, Slack, workspace

Visit Slack

How to install and configure Slack

1. Review the configuration options and authentication setup in the Slack MCP Server repository. 2. Choose between stealth mode or standard OAuth mode based on your workspace administrative requirements. 3. Configure your MCP client (such as Claude Desktop or Cursor) to register the server using your preferred transport (Stdio, SSE, or HTTP). 4. Enable write tools if needed by setting environment variables like SLACK_MCP_ADD_MESSAGE_TOOL and SLACK_MCP_REACTION_TOOL in your client's MCP configuration block. 5. Restart your MCP client to verify connectivity and access workspace tools.

What you can do with Slack

• Fetching recent discussions and thread replies from specific channels using channel names or channel IDs. - Locating historical decisions by searching messages filtered by sender, timestamp, channel, or direct message context. - Catching up on missed conversations by querying unread messages across channels sorted by mention priority. - Posting automated summaries or alerts directly into public channels or designated message threads when write permissions are enabled. - Adding or removing emoji reactions to manage message workflows and acknowledgement statuses directly from the AI client.

Key facts

• https://github.com/korotovsky/slack-mcp-server
• Communication & Messaging
• collaboration, messaging, Slack, workspace

Part of MCP Servers

What can Slack MCP do?

Slack MCP allows Model Context Protocol clients to inspect Slack workspaces. It can fetch channel histories, read message threads, retrieve channel directories, and execute filtered searches across public channels, private conversations, and direct messages. When explicitly enabled by environment variables, it can also post messages and manage emoji reactions.

Does Slack MCP require workspace administrator approval?

Not necessarily. The server supports a stealth mode that can operate without requiring bot installations or administrator approvals in the workspace. Alternatively, users can configure standard OAuth access when administrator-approved integrations are preferred.

Can Slack MCP post messages to channels?

Yes, but posting messages is disabled by default for safety. To enable message posting, you must set the SLACK_MCP_ADD_MESSAGE_TOOL environment variable, which can be configured globally or restricted to a comma-separated list of allowed channel IDs.

Which transports are supported by Slack MCP?

Slack MCP supports Stdio, Server-Sent Events (SSE), and HTTP transports. It also includes proxy support, allowing outgoing Slack API requests to be routed through an HTTP or enterprise network proxy if required.

Why does search fail when using a bot token with Slack MCP?

The search messages tool is not supported when using Slack bot tokens starting with xoxb. This is an API limitation imposed by Slack, as bot tokens are restricted from calling the Slack search.messages endpoint.