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 For Good AI Inc. (dba Zencoder Inc.), 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 search channels, post messages, reply to threads, add reactions, get channel history, and manage users.

Available Tools

1. slack_search_channels

   • Search for channels by name using the cached channel list
   • Supports partial, case-insensitive matching
   • Automatically strips # prefix from queries
   • Required inputs:
     • query (string): Channel name to search for (partial match, case-insensitive, # prefix optional)
   • Returns: List of matching channels with their IDs and information

2. slack_post_message

   • Post a new message to a Slack channel or direct message to user
   • Required inputs:
     • channel_id (string): The ID of the channel to post to
     • text (string): The message text to post
   • Returns: Message posting confirmation and timestamp

3. slack_reply_to_thread

   • Reply to a specific message thread
   • Required inputs:
     • channel_id (string): The channel containing the thread
     • thread_ts (string): Timestamp of the parent message in the format '1234567890.123456'. Timestamps without the period can be converted by adding the period such that 6 numbers come after it.
     • text (string): The reply text
   • Returns: Reply confirmation and timestamp

4. slack_add_reaction

   • Add an emoji reaction to a message
   • Required inputs:
     • channel_id (string): The channel containing the message
     • timestamp (string): Message timestamp to react to
     • reaction (string): Emoji name without colons
   • Returns: Reaction confirmation

5. slack_get_channel_history

   • Get messages from a channel within a time range
   • Default behavior: Retrieves last 24 hours of messages if no time range specified
   • Timestamp formats supported: ISO date strings, Unix timestamps, or Slack timestamps
   • Messages matching exact timestamps are included (inclusive behavior)
   • Required inputs:
     • channel_id (string): The channel ID
   • Optional inputs:
     • oldest (string | number): Start of time range. Accepts:
       • ISO date string (e.g., '2024-01-15T10:00:00Z')
       • Unix timestamp (e.g., 1609459200)
       • Slack timestamp (e.g., '1609459200.123456')
       • Defaults to 24 hours ago if not provided
     • latest (string | number): End of time range. Same formats as oldest. Defaults to now if not provided.
   • Returns: List of messages (up to 200) within the time range with their content and metadata
   • Examples:

     // Get last 24 hours (default)
     { channel_id: 'C123456' }
     
     // Get specific time range with ISO dates
     { channel_id: 'C123456', oldest: '2024-01-15T10:00:00Z', latest: '2024-01-15T18:00:00Z' }
     
     // Get specific time range with Unix timestamps
     { channel_id: 'C123456', oldest: 1609459200, latest: 1609545600 }
     
     // Get everything since a specific time (latest defaults to now)
     { channel_id: 'C123456', oldest: '2024-01-15T00:00:00Z' }
     

6. slack_get_thread_replies

   • Get all replies in a message thread
   • Required inputs:
     • channel_id (string): The channel containing the thread
     • thread_ts (string): Timestamp of the parent message in the format '1234567890.123456'. Timestamps without the period can be converted by adding the period such that 6 numbers come after it.
   • Returns: List of replies with their content and metadata

7. slack_get_users

   • Get list of workspace users with basic profile information
   • Optional inputs:
     • cursor (string): Pagination cursor for next page
     • limit (number, default: 100, max: 200): Maximum users to return
   • Returns: List of users with their basic profiles

8. 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

How Channel Search Works

The server initializes a local cache of all workspace channels on startup. The slack_search_channels tool performs fast, client-side searches against this cache rather than making API calls for each search. This provides:

• Faster search responses
• Partial name matching (e.g., searching "gen" will find "general")
• Case-insensitive matching
• Automatic # prefix handling

The cache is built once during server initialization by fetching all channels via pagination.

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:

Required scopes:

• channels:history - View messages and other content in public channels
• channels:read - View basic channel information
• chat:write - Send messages as the app
• reactions:write - Add emoji reactions to messages
• users:read - View users and their basic information
• users.profile:read - View detailed profiles about users

Additional scopes for private channel access:

• groups:history - View messages in private channels
• groups:read - View basic private channel information

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:
  • Search channels (with caching for fast partial-match searches)
  • Post messages
  • Reply to threads
  • 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 @benbroks/slack-mcp-server

Docker Installation

# Build the Docker image locally
docker build -t slack-mcp-server .

# Or pull from Docker Hub
docker pull hynzk6uuwdrdd9na/slack-mcp:latest

# Or pull a specific version
docker pull hynzk6uuwdrdd9na/slack-mcp:1.0.0

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

Docker Usage Examples

Using Docker directly

# Run with stdio transport (default)
docker run --rm \
  -e SLACK_BOT_TOKEN="xoxb-your-bot-token" \
  -e SLACK_TEAM_ID="your-team-id" \
  hynzk6uuwdrdd9na/slack-mcp:latest

# Run with HTTP transport on port 3000
docker run --rm -p 3000:3000 \
  -e SLACK_BOT_TOKEN="xoxb-your-bot-token" \
  -e SLACK_TEAM_ID="your-team-id" \
  hynzk6uuwdrdd9na/slack-mcp:latest --transport http

# Run with HTTP transport on custom port
docker run --rm -p 8080:8080 \
  -e SLACK_BOT_TOKEN="xoxb-your-bot-token" \
  -e SLACK_TEAM_ID="your-team-id" \
  hynzk6uuwdrdd9na/slack-mcp:latest --transport http --port 8080

# Run with custom auth token
docker run --rm -p 3000:3000 \
  -e SLACK_BOT_TOKEN="xoxb-your-bot-token" \
  -e SLACK_TEAM_ID="your-team-id" \
  -e AUTH_TOKEN="mytoken" \
  hynzk6uuwdrdd9na/slack-mcp:latest --transport http

Using Docker Compose

Create a docker-compose.yml file:

version: '3.8'

services:
  slack-mcp:
    # Use published image:
    image: hynzk6uuwdrdd9na/slack-mcp:latest
    # Or build locally:
    # build: .
    environment:
      - SLACK_BOT_TOKEN=xoxb-your-bot-token
      - SLACK_TEAM_ID=your-team-id
      - SLACK_CHANNEL_IDS=channel1,channel2,channel3  # Optional
      - AUTH_TOKEN=your-auth-token  # Optional for HTTP transport
    ports:
      - "3000:3000"  # Only needed for HTTP transport
    command: ["--transport", "http"]  # Optional: specify transport type
    restart: unless-stopped

Then run:

# Start the service
docker compose up -d

# View logs
docker compose logs -f slack-mcp

# Stop the service
docker compose down

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:

1. Command Line: Use --token <token> to specify a custom token
2. Environment Variable: Set AUTH_TOKEN=<token> as a fallback
3. 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:

1. All required scopes are added to your Slack app
2. The app is properly installed to your workspace
3. The tokens and workspace ID are correctly copied to your configuration
4. The app has been added to the channels it needs to access

Development

Build

npm run build

Watch Mode

npm run watch

API Endpoints (Streamable HTTP Transport)

When using Streamable HTTP transport, the server exposes the following endpoints:

• POST /mcp - Client-to-server communication
• GET /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