MCP Academy · 课时

面向客户端的输出模式

公开数据结构，让使用者可以据此依赖。

面向客户端的输出模式 是 CoddyKit 上的免费 MCP Academy 课时。这是第 3 节课，共 4 节。你可以在下方免费阅读本课时的完整内容 — 然后在浏览器中使用内置代码编辑器和全天候 AI 导师进行实践。这是 MCP Academy 学习路径的一部分，你的进度在网页和 CoddyKit 应用中同步。 MCP Academy 课程共包含 4 节课。

本课时的部分内容尚未翻译，以英文显示。

Publish the Shape

A structured tool can also advertise an output schema: a JSON Schema that tells clients exactly what fields and types to expect.

What an Output Schema Is

It is a machine-readable description of your result, listing each field name and its type. Clients read it before ever calling the tool.

Generated From Your Model

When you return a Pydantic model, the SDK derives the output schema for you. Your type hints become the published contract automatically.

Where Clients See It

The schema rides along in the tool listing under outputSchema. A client inspecting tools learns the shape without trial and error.

{"type": "object", "properties": {"temp_c": {"type": "number"}}}

Why Clients Want It

With a known schema, a client can validate results and build UI ahead of time. There is no guessing about which fields will appear.

Types Carry Meaning

The schema marks temp_c as a number and condition as a string. Consumers trust these types instead of inferring them from one sample.

Required vs Optional

Fields with defaults become optional in the schema; others are required. Clients learn which keys are guaranteed to be present.

class Weather(BaseModel):
    temp_c: float
    note: str = ""

A Stable Contract

Treat the output schema like a promise. Once clients depend on it, changing field names or types can break them downstream.

Schemas Aid Validation

Strict clients validate every result against the schema and reject mismatches. This catches server bugs fast, right at the boundary.

Self-Documenting Tools

An output schema doubles as documentation. New developers read the schema to understand a tool without digging through your code.

Keep Schemas Honest

Make sure your tool always returns what the schema claims. A result that drifts from its schema erodes the trust clients place in it.

Quick Check

Where does a tool publish the shape of its result?

Recap: Output Schemas

You saw how a tool publishes an output schema from your model. It documents fields, enables validation, and forms a stable contract. ✅

常见问题解答

「面向客户端的输出模式」课时是免费的吗？

是的 — 「面向客户端的输出模式」的完整文本可在网页上免费阅读。要进行交互式练习（内置代码编辑器和全天候 AI 导师）并解锁 MCP Academy 课程的其余内容，请升级到 CoddyKit PRO。 MCP Academy 课程共包含 4 节课。

「面向客户端的输出模式」这节课中我会学到什么？

公开数据结构，让使用者可以据此依赖。你通过在浏览器中直接运行的动手代码来练习 MCP Academy，全天候 AI 导师会在你学习这节课的过程中回答你的问题。

学习 MCP Academy 需要有经验吗？

无需任何先前经验。CoddyKit 上的 MCP Academy 课程适合初学者到高级学习者，你可以从这里开始或从头开始，按照自己的节奏学习。这是第 3 节课，共 4 节。

「面向客户端的输出模式」课时需要多长时间？

大多数 CoddyKit 课程大约需要 5–10 分钟。每节课都很精短且互动，所以你能稳步进步，并在网页和应用中从离开的地方继续。

我能在这节 MCP Academy 课中编写并运行代码吗？

能。每节 MCP Academy 课都包含内置代码编辑器，你可以在浏览器中直接编写并运行真实代码，并获得即时 AI 反馈 — 无需本地设置。

此课程中的所有课时

← 返回 MCP Academy