slack-mcp-server
Enables AI assistants to interact with Slack workspaces, including listing channels, posting messages, replying to threads, adding reactions, and retrieving user information.
README
slack-mcp-server
Disclaimer
This project includes code originally developed by Anthropic and released under the MIT License. Substantial modifications and new functionality have been added by dudztroyer, and are licensed under the Apache License, Version 2.0.
Overview
A Model Context Protocol (MCP) server for interacting with Slack workspaces. This server provides tools to list channels, post messages (with optional markdown formatting), reply to threads (with optional markdown formatting), add reactions, get channel history, get thread replies, and manage users.
Available Tools
-
slack_list_channels
- List public or pre-defined channels in the workspace
- Optional inputs:
limit(number, default: 100, max: 200): Maximum number of channels to returncursor(string): Pagination cursor for next page
- Returns: List of channels with their IDs and information
-
slack_post_message
- Post a new message to a Slack channel
- Required inputs:
channel_id(string): The ID of the channel to post totext(string): The message text to post
- Returns: Message posting confirmation and timestamp
-
slack_reply_to_thread
- Reply to a specific message thread
- Required inputs:
channel_id(string): The channel containing the threadthread_ts(string): Timestamp of the parent messagetext(string): The reply text
- Returns: Reply confirmation and timestamp
-
slack_add_reaction
- Add an emoji reaction to a message
- Required inputs:
channel_id(string): The channel containing the messagetimestamp(string): Message timestamp to react toreaction(string): Emoji name without colons
- Returns: Reaction confirmation
-
slack_get_channel_history
- Get recent messages from a channel
- Required inputs:
channel_id(string): The channel ID
- Optional inputs:
limit(number, default: 10): Number of messages to retrieve
- Returns: List of messages with their content and metadata
-
slack_get_thread_replies
- Get all replies in a message thread
- Required inputs:
channel_id(string): The channel containing the threadthread_ts(string): Timestamp of the parent message
- Returns: List of replies with their content and metadata
-
slack_get_users
- Get list of workspace users with basic profile information
- Optional inputs:
cursor(string): Pagination cursor for next pagelimit(number, default: 100, max: 200): Maximum users to return
- Returns: List of users with their basic profiles
-
slack_get_user_profile
- Get detailed profile information for a specific user
- Required inputs:
user_id(string): The user's ID
- Returns: Detailed user profile information
-
slack_post_markdown_message
- Post a new message to a Slack channel with markdown formatting enabled
- Required inputs:
channel_id(string): The ID of the channel or user to post totext(string): The message text with markdown formatting (mrkdwn). Supports bold, italic,code,code blocks, >quotes, <@user_id> mentions, <#channel_id> channel links, and <url|link text> links
- Returns: Message posting confirmation and timestamp
-
slack_reply_to_thread_markdown
- Reply to a specific message thread in Slack with markdown formatting enabled
- Required inputs:
channel_id(string): The ID of the channel containing the threadthread_ts(string): The timestamp of the parent message in the format '1234567890.123456'. Timestamps in the format without the period can be converted by adding the period such that 6 numbers come after ittext(string): The reply text with markdown formatting (mrkdwn). Supports bold, italic,code,code blocks, >quotes, <@user_id> mentions, <#channel_id> channel links, and <url|link text> links
- Returns: Reply confirmation and timestamp
Slack Bot Setup
To use this MCP server, you need to create a Slack app and configure it with the necessary permissions:
1. Create a Slack App
- Visit the Slack Apps page
- Click "Create New App"
- Choose "From scratch"
- Name your app and select your workspace
2. Configure Bot Token Scopes
Navigate to "OAuth & Permissions" and add these scopes:
channels:history- View messages and other content in public channelschannels:read- View basic channel informationchat:write- Send messages as the appreactions:write- Add emoji reactions to messagesusers:read- View users and their basic informationusers.profile:read- View detailed profiles about users
3. Install App to Workspace
- Click "Install to Workspace" and authorize the app
- Save the "Bot User OAuth Token" that starts with
xoxb-
4. Get Your Team ID
Get your Team ID (starts with a T) by following this guidance
5. Add Bot to Channels (Optional)
For the bot to access private channels or to post messages, you may need to invite it to specific channels using /invite @your-bot-name
Features
- Multiple Transport Support: Supports both stdio and Streamable HTTP transports
- Modern MCP SDK: Updated to use the latest MCP SDK (v1.13.2) with modern APIs
- Comprehensive Slack Integration: Full set of Slack operations including:
- List channels (with predefined channel support)
- Post messages (plain text and markdown formatting)
- Reply to threads (plain text and markdown formatting)
- Add reactions
- Get channel history
- Get thread replies
- List users
- Get user profiles
Installation
Local Development
npm install
npm run build
Global Installation (NPM)
npm install -g dudztroyer/slack-mcp-server
Configuration
Set the following environment variables:
export SLACK_BOT_TOKEN="xoxb-your-bot-token"
export SLACK_TEAM_ID="your-team-id"
export SLACK_CHANNEL_IDS="channel1,channel2,channel3" # Optional: predefined channels
export AUTH_TOKEN="your-auth-token" # Optional: Bearer token for HTTP authorization (Streamable HTTP transport only)
Usage
Command Line Options
slack-mcp [options]
Options:
--transport <type> Transport type: 'stdio' or 'http' (default: stdio)
--port <number> Port for HTTP server when using Streamable HTTP transport (default: 3000)
--token <token> Bearer token for HTTP authorization (optional, can also use AUTH_TOKEN env var)
--help, -h Show this help message
Local Usage Examples
Using the slack-mcp command (after global installation)
# Use stdio transport (default)
slack-mcp
# Use stdio transport explicitly
slack-mcp --transport stdio
# Use Streamable HTTP transport on default port 3000
slack-mcp --transport http
# Use Streamable HTTP transport on custom port
slack-mcp --transport http --port 8080
# Use Streamable HTTP transport with custom auth token
slack-mcp --transport http --token mytoken
# Use Streamable HTTP transport with auth token from environment variable
AUTH_TOKEN=mytoken slack-mcp --transport http
Using node directly (for development)
# Use stdio transport (default)
node dist/index.js
# Use stdio transport explicitly
node dist/index.js --transport stdio
# Use Streamable HTTP transport on default port 3000
node dist/index.js --transport http
# Use Streamable HTTP transport on custom port
node dist/index.js --transport http --port 8080
# Use Streamable HTTP transport with custom auth token
node dist/index.js --transport http --token mytoken
# Use Streamable HTTP transport with auth token from environment variable
AUTH_TOKEN=mytoken node dist/index.js --transport http
Transport Types
Stdio Transport
- Use case: Command-line tools and direct integrations
- Communication: Standard input/output streams
- Default: Yes
Streamable HTTP Transport
- Use case: Remote servers and web-based integrations
- Communication: HTTP POST requests with optional Server-Sent Events streams
- Features:
- Session management
- Bidirectional communication
- Resumable connections
- RESTful API endpoints
- Bearer token authentication
Authentication (Streamable HTTP Transport Only)
When using Streamable HTTP transport, the server supports Bearer token authentication:
- Command Line: Use
--token <token>to specify a custom token - Environment Variable: Set
AUTH_TOKEN=<token>as a fallback - Auto-generated: If neither is provided, a random token is generated
The command line option takes precedence over the environment variable. Include the token in HTTP requests using the Authorization: Bearer <token> header.
Troubleshooting
If you encounter permission errors, verify that:
- All required scopes are added to your Slack app
- The app is properly installed to your workspace
- The tokens and workspace ID are correctly copied to your configuration
- The app has been added to the channels it needs to access
Development
Build
npm run build
Watch Mode
npm run watch
Creating Releases
This project includes an automated release workflow that can be triggered manually via GitHub Actions:
How to Create a Release
- Navigate to GitHub Actions: Go to the "Actions" tab in your GitHub repository
- Select Release Workflow: Click on "Create Release" workflow
- Run Workflow: Click "Run workflow" button
- Configure Release:
- Version bump type: Choose from
patch,minor, ormajor
- Version bump type: Choose from
What the Release Workflow Does
The automated release process leverages changelogen to handle most of the release process:
- Quality Checks: Ensures all tests pass and code builds successfully
- Changelogen Release: Uses
changelogen --release --pushandchangelogen gh releaseto:- Generate beautiful changelog from Conventional Commits
- Bump version in
package.jsonaccording to your selection - Create and push git tag
- Commit and push all changes
- Create GitHub release with changelog as release notes
- Distribution & Assets: Adds distribution archive to the GitHub release
- Trigger Downstream: Automatically triggers:
- Package publishing to GitHub Package Registry (
publish.yml)
- Package publishing to GitHub Package Registry (
Version Bump Types
- Patch (
1.0.0→1.0.1): Bug fixes and minor changes - Minor (
1.0.0→1.1.0): New features that don't break existing functionality - Major (
1.0.0→2.0.0): Breaking changes
Conventional Commits for Better Changelogs
The release workflow uses changelogen which works best with Conventional Commits. To get the most beautiful and informative changelogs, format your commit messages like:
feat: add new slack reaction tool- New featuresfix: resolve channel history pagination issue- Bug fixesdocs: update README with Docker instructions- Documentation changeschore: update dependencies- Maintenance tasksrefactor: simplify error handling logic- Code refactoringperf: optimize message processing- Performance improvements
This will result in properly categorized and formatted changelog entries.
API Endpoints (Streamable HTTP Transport)
When using Streamable HTTP transport, the server exposes the following endpoints:
POST /mcp- Client-to-server communicationGET /mcp- Server-to-client notifications (Server-Sent Events streams)DELETE /mcp- Session termination
Changes from Previous Version
- Updated MCP SDK: Upgraded from v1.0.1 to v1.13.2
- Modern API: Migrated from low-level Server class to high-level McpServer class
- Zod Validation: Added proper schema validation using Zod
- Transport Flexibility: Added support for Streamable HTTP transport
- Command Line Interface: Added CLI arguments for transport selection
- Session Management: Implemented proper session handling for HTTP transport
- Better Error Handling: Improved error handling and logging
Recommended Servers
playwright-mcp
A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.
Magic Component Platform (MCP)
An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.
Audiense Insights MCP Server
Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.
VeyraX MCP
Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.
graphlit-mcp-server
The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.
Kagi MCP Server
An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.
Neon Database
MCP server for interacting with Neon Management API and databases
E2B
Using MCP to run code via e2b.
Exa Search
A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.
Qdrant Server
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.