Whatever message this page gives is out now! Go check it out!
Purpose of MCP and its role in ColdFusion
Last update:
May 18, 2026
Purpose of MCP
As AI systems evolve from simple chatbots to agents capable of complex tasks, they need a reliable way to access tools and data. The Model Context Protocol (MCP) is the protocol that AI agents (clients) use to communicate with external systems (servers) in a consistent, auditable way.
MCP standardizes:
- Tool discovery and invocation:
tools/listandtools/call - Prompt management:
prompts/listandprompts/get - Resource access:
resources/listandresources/read - Optional flows: Sampling, elicitation, logging, and roots
MCP in ColdFusion
ColdFusion supports MCP in two primary ways:
- Call MCP servers: Your ColdFusion application can act as an MCP client and call external or internal MCP servers using a structured client API.
- Expose an MCP server: Your ColdFusion business logic, prompts, and resources can be exposed as an MCP server that any MCP host can connect to.
Key components
MCP is built around three primary roles. ColdFusion can fulfill more than one of these roles.
MCP Host
The MCP Host is the outer application that orchestrates the conversation and user experience. It embeds an MCP client and typically an LLM (large language model).
Examples: Claude Desktop, Cursor, and ColdFusion web applications that maintain user sessions and call MCP tools as needed.
Responsibilities:
- Manage the user interface and conversation flow
- Decide when to call tools, fetch resources, or retrieve prompts
- Forward sampling and elicitation requests to the appropriate handlers
MCP Client
The MCP Client is the protocol adapter that communicates with MCP servers over HTTP or STDIO. It translates between MCP JSON-RPC requests and responses and ColdFusion structures.
Entry point:
MCPClient(configData) is the primary way to create MCP clients in ColdFusion.Responsibilities:
- Perform the initialization handshake with MCP servers
- Expose CF-friendly methods such as
listTools(),callTool(),listResources(),readResource(),listPrompts(), andgetPrompt() - Optionally handle sampling, elicitation, and logging callbacks
MCP Server
The MCP Server is the provider of capabilities. A server may wrap external APIs, a file system, or ColdFusion CFCs. It receives MCP methods such as
tools/list, tools/call, prompts/get, and returns standardized responses. Entry point:
MCPServer(configData) creates an MCP server in ColdFusion.Responsibilities:
- Expose tools (CFC methods), prompts, and resources
- Execute tool calls and return structured results
- Serve resources and prompt templates to clients
Client versus server in ColdFusion
In ColdFusion, the distinction between client and server is straightforward.
Use the MCP client
Use the MCP Client when your ColdFusion application needs to call external MCP servers.
Examples of external servers:
- Jira MCP
- Wiki MCP
- ColdFusion-based MCP servers
- Other third-party MCP servers
Typical flow:
- Create an
MCPClientwith transport configuration (HTTP or STDIO). - Call
listTools()to discover available tools. - Call
callTool()to invoke tools as needed. - Optionally use
listResources(),readResource(),listPrompts(), andgetPrompt().
Use the MCP server
Use the MCP Server when your ColdFusion application should act as an MCP server so that agents and tools can invoke your ColdFusion logic and access your data.
Typical flow:
- Create an
MCPServerwithserverInfo,capabilities, and tool/prompt/resource definitions. - Register CFCs as tools, define prompts, and configure resources.
- Expose an HTTP endpoint that delegates to
application.mcpServer.handleRequest(). - External MCP clients (including other ColdFusion apps) connect to your server.
Use both
Many teams use both approaches: they integrate existing MCP servers on the client side and expose their own ColdFusion capabilities via MCP on the server side.
We're glad. Tell us how this page helped.
We're sorry. Can you tell us what didn't work for you?
Share additional feedback to help us improve.
0/255
|
