Slack launched a platform refresh enabling third-party AI agents to integrate directly via a new Real-Time Search (RTS) API and Model Context Protocol (MCP) server. The RTS API gives agents live access to conversational data, files, and context within Slack channels and direct messages while respecting existing access permissions. The MCP server, based on an open-source framework developed by Anthropic, standardizes how agents access information and execute tasks across Slack, eliminating the need for developers to hand-code integrations for each LLM vendor.
Third-party agents from Anthropic (Claude), Perplexity, Google, and Writer are now available in the Slack Marketplace with simplified one-click installation via an "Add to Slack" button. Slack conducts quality and security reviews before publishing agents to the Marketplace, with explicit prohibitions on using Slack data to train LLMs. Workspace owners and admins retain control over which agent apps can be installed. The MCP server replaces fragmented integrations with a single protocol, reducing developer effort to connect agents to Slack significantly.
Slack also introduced Work Objects, which standardize how third-party data is displayed in Slack via dynamic, interactive previews. This allows actions like updating Asana tasks or fetching CRM data without app-switching. The platform enhancements are paired with new Slack AI native features: AI-powered search now surface files and apps, huddle transcription and note-taking, and an AI-driven Workflow Builder. Agentforce (Salesforce's custom agent builder) will integrate into Slack in early 2026.
For practitioners building on Slack: the RTS API and MCP server lower the barrier to deploying agents at organizational scale. The shift from app-specific integrations to a unified protocol matches Anthropic's broader Model Context Protocol vision and signals a move toward standardized LLM-tool bindings. Teams can now embed Claude, Perplexity, or custom agents directly into workflows without manual infrastructure setup. MCP server adoption across vendors could accelerate agent interoperability across platforms beyond Slack.