Slackdump provides an MCP server that lets AI agents and tools interact with Slack export archives through standard MCP transports (stdio or HTTP). The server exposes a set of tools to inspect and query Slack export data, such as listing channels and users, reading messages and threads, and retrieving workspace information. This enables AI agents to navigate your Slack data, perform targeted extractions, or prepare structured outputs for use in downstream tasks. You can run the built-in Slackdump MCP server and connect to it from your agent via the MCP protocol, then issue tools like load_source, list_channels, get_messages, and get_workspace_info to extract or inspect data.

To use it, install and start Slackdump, then initialize an MCP project with slackdump mcp -new <project-path>. The server will automatically provide the bundled Slackdump skill/instruction files so the agent knows how to interpret your archive. You can choose between stdio (agent-managed) or HTTP transports depending on your integration setup. The available tools include: load_source (open an archive), list_channels (enumerate channels), get_channel (channel details), list_users (members), get_messages (channel messages with pagination), get_thread (thread replies), get_workspace_info (workspace metadata), and command_help (CLI flag help for subcommands).

If you’re integrating with AI agents, you can point the agent at the MCP server and issue these tools to build structured queries against your Slack data, enabling tasks like summarization, search, or data extraction within your automation workflow.

Prerequisites:

• Go 1.20+ (or a compatible Go toolchain)
• Git
• Access to the internet to fetch dependencies

Install and run Slackdump MCP server from source (example build path may vary by version):

1. Clone the repository git clone https://github.com/rusq/slackdump.git cd slackdump

2. Build the Slackdump executable (example for v4 project structure) go build -o slackdump ./cmd/slackdump

3. Verify the binary runs ./slackdump help

4. Initialize an MCP project for Slackdump (one-time setup) ./slackdump mcp -new ~/my-slack-project

5. Start the MCP server via the MCP configuration (as defined in mcp_config)

   Typical invocation depending on your MCP orchestrator

   slackdump mcp -new ~/my-slack-project

6. Connect your MCP client (AI agent) to the server using the provided transport (stdio or HTTP) as configured in your environment.

Tips and notes:

• Slackdump requires appropriate Slack export data; some actions may require particular export formats (standard exports, Mattermost-compatible dumps, or Slackdump-specific archives).
• The MCP server supports both stdio and HTTP transports. Choose the transport that matches your agent’s capabilities.
• If you encounter permission or network issues when using the MCP server, ensure the environment variables and file paths in the mcp_config are correct and accessible by the process.
• Keep Slackdump up to date with the latest release to benefit from new tools and bug fixes.
• The Slackdump CLI provides a rich set of commands; use slackdump help and slackdump help <tool> to discover latest capabilities and flags.