Skip to main content
Bring your team and maximize your impact at Dreamforce. Register three or more to unlock $999 passes.

Explore Headless Tools and Functionality

Learning Objectives

After completing this unit, you’ll be able to:

  • Explain key tools and functionality that drive the Salesforce headless strategy.
  • Learn about the Salesforce support for Model Context Protocol (MCP) functionality.

Key Salesforce Tools and Functionality

The Salesforce headless strategy decouples the Salesforce Platform from any fixed UI or browser and makes every layer of the platform invokable for agents, models, and third-party systems. This strategy makes the platform accessible through any integrated development environment (IDE), model of choice, and tool.

Typically Salesforce customers access their orgs, data, and metadata through browser-based interfaces. With the Salesforce headless experience, customers can access Salesforce with these key tools:

  • Model Context Protocol (MCP)
  • Application programming interface (API)
  • Command-line interface (CLI)
  • Skills

These tools are used in IDEs such as Claude, Cursor, ChatGPT and Saleforce’s own integrated Agentforce Vibes IDE or Slack. These tools provide a wide range of benefits to developers, AI applications, or AI agents. In this badge, explore how the tools power headless capabilities. Learn how to enable and use the tools to build, integrate, and automate in whatever AI interface you choose.

Model Context Protocol

MCP is an open-source standard for connecting AI applications to external systems. MCP provides a standardized way for any agent to connect, call tools, query data, and take action. It’s a universal connector. Instead of building custom API integrations for each AI tool your team uses, you configure an MCP server once in Salesforce. And any compatible client such as Claude, Cursor, or Agentforce Vibes can connect to it by using standard OAuth-based authentication.

These are some of the MCP servers available to Salesforce customers.

  • Salesforce-hosted Standard MCP servers: These are standard MCP servers, bundling out-of-the-box platform capabilities. A Salesforce-hosted MCP server authenticates via Salesforce org. And it gives external AI agents access to your Salesforce metadata, data, and functionality, using MCP servers such as the Salesforce Headless 360 MCP or the Metadata Grounding MCP.
  • Salesforce-hosted product MCP servers: These are Salesforce product-specific MCP servers available through prebuilt integrations in hosted MCP servers. The MCP servers have product-specific server URLs. For example, Tableau Next MCP or MuleSoft MCP.
  • External MCP servers: These third-party servers require manual registration, installation, and they appear as external servers under API Catalog.
  • Custom MCP servers: By combining MCP tools from different servers to accomplish broader workflows, administrators can create and curate custom MCP servers. Grouping the tools a specific persona needs into one server URL also simplifies interactions with MCP clients—they only need a single URL to access all the relevant Salesforce tools. Custom servers also let you expose your own tools, backed by Apex code, Flows, or APIs, since standard servers can't be modified.

Let’s understand the functionality that Salesforce-hosted MCP servers provide.

Salesforce provides a set of hosted MCP servers that expose built-in platform and product capabilities. These servers are preconfigured by Salesforce and provide a fixed set of tools. These servers are disabled by default and can be enabled by an administrator in Salesforce, but their tool sets cannot be modified.

All servers enforce per-user authentication and respect the Salesforce standard security model: field-level security, object permissions, and sharing rules apply to every tool call. External AI agents can securely connect with your Salesforce data, assets, and logic safely within the Salesforce security perimeter and governance policies.

Any compatible AI client—Claude, ChatGPT, Cursor, or a custom agent—can connect to a Salesforce org and act on behalf of authorized users.

Examples of Salesforce-hosted MCP servers include Salesforce Headless 360 MCP Server (Beta), Data 360 MCP server, and Tableau Next MCP server. To learn more about the capability of each of these servers, check out Salesforce Hosted MCP Servers: Standard MCP Servers Reference.

Let’s dig into the Salesforce Headless 360 MCP server—the server that unlocks the entire breadth of Salesforce through one connection and just four tools.

Salesforce Headless 360 MCP Server (Beta)

The Headless 360 MCP Server (Beta) brings the full breadth of Salesforce to any MCP-aware agent through a single connection. The server exposes four tools that map to how an agent moves through a request, backed by a continuously growing library of Salesforce operations.

  • Discover: Finds the actions your agent can take.
  • Describe: Returns the full specification for an action.
  • Dispatch: Runs the chosen action.
  • Dispatch (Read-Only): Runs read-only actions.

This MCP Server design keeps your agent’s tool surface small and stable, while the set of actions it can take scales independently with every release. At beta launch, the server offers dozens of operations focused mostly on admin Setup tasks, with more added over time to cover what admins, developers, and business users do across Salesforce. For example, with it, an agent can query, create, and update records, manage users and permission sets, and set up named credentials.

And because it works across all of Salesforce, it’s your agent’s gateway to acting on the entire platform through one endpoint. For more information see, Headless 360 MCP Server (Beta).

What’s Next

In this unit, you explored the foundation of the Salesforce headless strategy. This approach decouples the platform from any fixed UI through four key tools: MCP, APIs, CLIs, and Skills. These tools are all supported across a broad ecosystem that includes Claude, Cursor, ChatGPT, Agentforce Vibes, and Slack. You took a closer look at MCP and the different types of MCP servers available to Salesforce Customers.

MCP gives AI agents a standardized, governed way to connect to Salesforce—but it’s just one part of the headless toolkit. In the next unit, learn how APIs provide the flexible foundation for programmatic integration and how scriptable CLIs give a command-line option for development and deployment. And explore how Skills package repeatable workflows into single commands that you can run in your AI interface.

Resources

Share your Trailhead feedback over on Salesforce Help.

We'd love to hear about your experience with Trailhead - you can now access the new feedback form anytime from the Salesforce Help site.

Learn More Continue to Share Feedback