Should the MCP Tool Metadata Include an 'Output_Schema' for Improved Workflow Integration?

Help
1

This topic was automatically generated from Slack. You can find the original thread here.

Hey Guys,

Currently, the MCP tool metadata includes only the input_schema, but there’s no corresponding output_schema provided in the response.
Having access to the output_schema would be extremely useful for workflow integration scenarios. It would allow us to:
• Automatically parse and map tool outputs to subsequent nodes in a workflow.
• Build stronger type validation and data handling between interconnected tools.
• Improve overall automation and reduce manual JSON parsing in custom integrations.
Proposed Solution:
Include an output_schema field in the MCP tool definition or metadata response, describing the expected structure of the tool’s output (e.g., JSON Schema).
Example:


{
  "name": "get_events",
  "input_schema": { ... },
  "output_schema": {
    "type": "object",
    "properties": {
      "events": {
        "type": "array",
        "items": {
          "type": "object",
          "properties": {
            "id": { "type": "string" },
            "title": { "type": "string" },
            "start_time": { "type": "string", "format": "date-time" }
          }
        }
      }
    }
  }
}

Use Case:
In our workflow platform, MCP tools are represented as nodes. With output_schema, we could automatically join and parse the output of one node into another tool, enabling seamless data flow between connected components.
Benefit:
This enhancement would make the MCP ecosystem more interoperable and developer-friendly, especially for platforms integrating multiple tool-based workflows.

Ref : Changelog - FastMCP

2

Thanks for the suggestion, this is something we have in mind and hopefully will be able to add in the future.

3

Thanks !