Salesforce Hosted MCP Servers are available
Salesforce confirmed this week that its Hosted MCP Servers have reached general availability, marking a production-ready milestone for the Model Context Protocol (MCP) capability it first piloted in Spring 2025 and launched into beta last October. For teams running Marketing Cloud Engagement, this GA announcement opens up a significant new channel for AI-powered campaign management and subscriber workflows.
What Are Salesforce Hosted MCP Servers?
MCP – originally developed by Anthropic – has emerged as the de facto open standard for giving AI agents structured, secure access to external systems.
MCP (Model Context Protocol) is a standard that lets AI models connect to external tools, data, and systems in a unified way.
A Salesforce Hosted MCP Server is a Salesforce-managed endpoint that exposes your org’s data, flows, Apex actions, and queries to any AI client that speaks MCP – including Claude, ChatGPT, and custom Agentforce agents. Salesforce handles the hosting, scaling, and uptime automatically, much like it does for the REST APIs already in use across most orgs. You enable a server in Setup and you’re live – no infrastructure to provision.
Secure by Design
One of the standout aspects of the GA release is how permissions are handled. Every MCP transaction runs with the authenticated user’s identity, meaning existing CRUD, field-level security (FLS), and sharing rules all apply automatically. If an AI agent updates a record, the human user’s name still appears in the audit trail. If their permissions don’t allow an operation, neither does the agent. For marketers managing sensitive subscriber data and data extensions across regions, that level of governance accountability matters.
Marketing Cloud Engagement Gets an AI Gateway
The GA release is particularly meaningful for Marketing Cloud Engagement users because Salesforce has also made an SFMC-specific MCP server available. This server connects AI tools directly to Marketing Cloud for campaign orchestration, content generation, and subscriber management – enabling a new class of marketing automation where AI agents can draft journeys, analyze audience segments, and manage send schedules without requiring a developer to write API scripts from scratch.
Teams that have already configured data syncs between Salesforce CRM and Marketing Cloud will find MCP servers a natural extension – the same CRM data surfaces can now be queried and acted upon by AI agents in real time.
Available Across Enterprise Editions – and Now Developer Edition Too
Access to Salesforce Hosted MCP Servers is included for every Enterprise Edition org and above at no additional cost. In a separate but related announcement on the Salesforce Developers Blog, Salesforce confirmed that Developer Edition orgs now include Hosted MCP Servers alongside the new Agentforce Vibes IDE and Claude Sonnet 4.5 as the default coding model – all at no cost. That is a meaningful signal for marketers and admins who want to prototype AI-assisted workflows before rolling them out in production.
What This Means in Practice
For Marketing Cloud practitioners, Salesforce Hosted MCP Servers moving to GA means AI agents can be given scoped, permission-controlled access to your Marketing Cloud org without custom middleware. Campaign assembly, audience segmentation, and content personalization tasks can be delegated to agentic AI with a full audit trail maintained. Third-party AI tools – Claude, ChatGPT, and others – can connect to SFMC without requiring dedicated API integration work, and the same governance controls marketers already rely on extend automatically to all AI-driven actions.
The official announcement confirms GA availability is rolling out now across eligible orgs, and with the Summer ’26 release cycle already underway, further Marketing Cloud Engagement enhancements tied to the MCP ecosystem are expected in the months ahead.
Share With Others
