ling LLM models to communicate with various MCP servers through standardized MCP tools. Important distinction: This client is not designed to interact with the Slack API directly as its primary purpose. However, it can achieve Slack API functionality by connecting to a dedicated Slack MCP server (such as [modelcontextprotocol/servers/slack](https://github.com/modelcontextprotocol/servers/tree/main/src/slack)) alongside other MCP servers. ## How It Works ![Image](https://github.com/user-attachments/assets/48a587e4-7895-4a6f-9745-61b21894c34c) ```mermaid flowchart LR User([User]) --> SlackBot subgraph SlackBotService[Slack Bot Service] SlackBot[Slack Bot] <--> MCPClient[MCP Client] end MCPClient <--> MCPServer1[MCP Server 1] MCPClient <--> MCPServer2[MCP Server 2] MCPClient <--> MCPServer3[MCP Server 3] MCPServer1 <--> Tools1[(Tools)] MCPServer2 <--> Tools2[(Tools)] MCPServer3 <--> Tools3[(Tools)] style SlackBotService fill:#F8F9F9,stroke:#333,stroke-width:2px style SlackBot fill:#F4D03F,stroke:#333,stroke-width:2px style MCPClient fill:#2ECC71,stroke:#333,stroke-width:2px style MCPServer1 fill:#E74C3C,stroke:#333,stroke-width:2px style MCPServer2 fill:#E74C3C,stroke:#333,stroke-width:2px style MCPServer3 fill:#E74C3C,stroke:#333,stroke-width:2px style Tools1 fill:#9B59B6,stroke:#333,stroke-width:2px style Tools2 fill:#9B59B6,stroke:#333,stroke-width:2px style Tools3 fill:#9B59B6,stroke:#333,stroke-width:2px ``` 1. **User** interacts only with Slack, sending messages through the Slack interface 2. **Slack Bot Service** is a single process that includes: - The Slack Bot component that handles Slack messages - The MCP Client component that communicates with MCP servers 3. The **MCP Client** forwards requests to the appropriate MCP Server(s) 4. **MCP Servers** execute their respective tools and return results ## Features - ✅ **Multi-Mode MCP Client**: - Server-Sent Events (SSE) for real-time communication - HTTP transport for JSON-RPC - stdio for local development and testing - ✅ **Slack Integration**: - Uses Socket Mode for secure, firewall-friendly communication - Works with both channels and direct messages - Rich message formatting with Markdown and Block Kit - Automatic conversion of quoted strings to code blocks for better readability - ✅ **LLM Support**: - Native OpenAI integration - LangChain gateway for multiple LLM providers - ✅ **Tool Registration**: Dynamically register and call MCP tools - ✅ **Docker container support** ## Installation #### From Source ### Running Locally with Binary After installing the binary, you can run it locally with the following steps: 1. Set up environment variables: ```bash # Using environment variables directly export SLACK_BOT_TOKEN="xoxb-your-bot-token" export SLACK_APP_TOKEN="xapp-your-app-token" export OPENAI_API_KEY="sk-your-openai-key" export OPENAI_MODEL="gpt-4o" export LOG_LEVEL="info" # Or create a .env file and source it cat > .env << EOL SLACK_BOT_TOKEN="xoxb-your-bot-token" SLACK_APP_TOKEN="xapp-your-app-token" OPENAI_API_KEY="sk-your-openai-key" OPENAI_MODEL="gpt-4o" LOG_LEVEL="info" EOL source .env ``` 2. Create an MCP servers configuration file: ```bash # Create mcp-servers.json in the current directory cat > mcp-servers.json << EOL { "mcpServers": { "filesystem": { "command": "npx", "args": ["-y", "@modelcontextprotocol/server-filesystem", "$HOME"], "env": {} } } } EOL ``` 3. Run the application: ```bash # Run with default settings (looks for mcp-servers.json in current directory) slack-mcp-client # Or specify a custom config file location slack-mcp-client --config /path/to/mcp-servers.json # Enable debug mode slack-mcp-client --debug # Specify OpenAI model slack-mcp-client --openai-model gpt-4o-mini ``` The application will connect to Slack and start listening for messages. You can check the logs for any errors or connection issues. ### Kubernetes Deployment with Helm For de

Slack MCP Client in Go

Comments