Slack MCP Server is an open-source Model Context Protocol server that gives AI assistants deep access to Slack workspaces — reading and posting messages, browsing channels, and fetching conversation history — without requiring a formally installed Slack app. It supports standard MCP transports and both OAuth and a browser-token stealth mode. It is an open-source project on GitHub with around 1,900 stars, actively developed through 2026 and used by tens of thousands of engineers monthly according to its maintainers.
Connecting an LLM agent to Slack normally means creating a Slack app, negotiating scopes with workspace admins and living within strict API tiers. This server offers two paths: a conventional OAuth mode, and a stealth mode that works with user session tokens so individuals can wire their own assistant into Slack with no admin approval — including support for government Slack (GovSlack) and enterprise workspaces. Its smart history fetching retrieves conversations by date or message count across channels, DMs and group DMs, which makes it practical for summarization, search and support-triage agents.
Free. Slack MCP Server is open source under the MIT license; you run it yourself and use your own Slack account.
Developers and power users building Slack-connected agents in Claude, Cursor or other MCP clients, especially where installing an official Slack app is impractical.