Add a pluggable server extension API with MCP Apps (#3003)

Co-authored-by: Max Isbey <224885523+maxisbey@users.noreply.github.com>

Author: Kludex

docs/advanced/apps.md

@@ -0,0 +1,160 @@
+# MCP Apps
+
+An **MCP App** is a tool with a face: alongside its data, the tool points at an HTML
+document the host renders as an interactive surface.
+
+Two parts, always two parts:
+
+1. **A tool** that does the work and returns data, like any other tool.
+2. **A `ui://` resource** containing the HTML the host shows for it.
+
+The tool carries a `_meta.ui.resourceUri` reference to the resource. The host fetches
+it with `resources/read`, renders it in a **sandboxed iframe**, and pushes the tool's
+result into that iframe via `postMessage`. Your server never sends or receives any
+`ui/*` messages: that traffic is between the host and the iframe. You serve a tool
+and an HTML document; the host does the theater.
+
+The SDK ships this as the built-in `Apps` extension (`io.modelcontextprotocol/ui`).
+If [Extensions](extensions.md) are new to you, skim that page first. One minute,
+then come back.
+
+## A clock with a face
+
+```python title="server.py" hl_lines="18 21 29 31"
+--8<-- "docs_src/apps/tutorial001.py"
+```
+
+Four moves:
+
+* `Apps()`: one instance holds your UI-bound tools and their resources.
+* `@apps.tool(resource_uri="ui://clock/app.html")`: a regular tool, plus the
+  `_meta.ui.resourceUri` stamp. Everything `@mcp.tool()` accepts (name, title,
+  description, ...) passes through.
+* `apps.add_html_resource("ui://clock/app.html", CLOCK_HTML)`: the matching
+  resource, served as `text/html;profile=mcp-app`. That exact MIME type is what
+  tells a host "this is an app, render it".
+* `MCPServer("clock", extensions=[apps])`: opt in. The server now advertises
+  `io.modelcontextprotocol/ui` under `capabilities.extensions`.
+
+The HTML itself listens for the host's `postMessage` and shows the result. For real
+apps, use the official [`@modelcontextprotocol/ext-apps`](https://github.com/modelcontextprotocol/ext-apps)
+browser SDK inside your HTML. It gives you `ontoolresult`, `callServerTool`,
+`getHostContext`, and `onhostcontextchanged` instead of raw message events.
+
+## Graceful degradation
+
+Not every client renders apps. The spec is blunt about what that means for you:
+
+> Tools **MUST** return a meaningful `content` array even when UI is available.
+
+The model reads `content`; the iframe is for humans. A UI-capable host still feeds
+the text result to the model, and a text-only client gets *only* that. So the
+canonical pattern is one tool, two answers. Look at `get_time` again:
+
+```python title="server.py" hl_lines="22-26"
+--8<-- "docs_src/apps/tutorial001.py"
+```
+
+`client_supports_apps(ctx)` is `True` only when the client declared the
+`io.modelcontextprotocol/ui` extension **and** listed `text/html;profile=mcp-app`
+in its `mimeTypes` settings. The field is required, so a client that omits it
+does not count. That is exactly what `main()` in the same file declares: the
+client half of the negotiation, and the rich answer comes back.
+
+!!! warning
+    Never return a placeholder like `"[Rendered UI]"` as the only content. If the
+    fallback text is useless, the tool is useless to every text-only client and to
+    the model itself. Write the sentence.
+
+## Locking the iframe down
+
+The resource side carries the security metadata: what the iframe may load, which
+browser permissions it wants, how it would like to be framed:
+
+```python title="server.py" hl_lines="9 19-22"
+--8<-- "docs_src/apps/tutorial002.py"
+```
+
+`csp` and `permissions` are **requests to the host**, not server behaviour. The host
+builds the iframe's Content-Security-Policy and Permissions-Policy from them, and it
+may refuse. Feature-detect in your JS rather than assuming a grant.
+
+`ResourceCsp`, field by field (Python name, wire key, what the host does with it):
+
+| Python | Wire (`_meta.ui.csp`) | Controls |
+|---|---|---|
+| `connect_domains` | `connectDomains` | `connect-src`: where `fetch`/XHR may go |
+| `resource_domains` | `resourceDomains` | `img-src`, `style-src`, ...: static assets |
+| `frame_domains` | `frameDomains` | `frame-src`: nested iframes |
+| `base_uri_domains` | `baseUriDomains` | `base-uri`: what `<base>` may point at |
+
+`ResourcePermissions`: each field requests a browser permission for the iframe.
+
+| Python | Wire (`_meta.ui.permissions`) |
+|---|---|
+| `camera` | `camera` |
+| `microphone` | `microphone` |
+| `geolocation` | `geolocation` |
+| `clipboard_write` | `clipboardWrite` |
+
+!!! note
+    CSP and permissions live on the **resource**, never on the tool. The spec's tool
+    metadata has no slot for them, and hosts ignore them there. The SDK makes the
+    mistake unrepresentable: `@apps.tool()` simply has no `csp` parameter.
+
+### Visibility
+
+`visibility=["app"]` on a tool says "this exists for the iframe, not the model":
+
+* `"model"`: the model may call it.
+* `"app"`: the iframe may call it (via `callServerTool`).
+* Omitted: both, which is the default.
+
+Filtering is the **host's** job. Your server lists app-only tools in `tools/list`
+like any other; the host hides them from the model. Don't filter server-side.
+
+## The rules the SDK enforces
+
+All of these fail at startup, not in production:
+
+* A `resource_uri` or resource URI that isn't `ui://...` is a `ValueError` at
+  decoration/registration time.
+* A tool bound to a URI with **no matching registered resource** is a `ValueError`
+  when `MCPServer(extensions=[apps])` consumes the extension. A tool advertising
+  HTML that 404s on `resources/read` is a misconfiguration, so it refuses to
+  construct.
+* `meta={"ui": ...}` on `@apps.tool()` is a `ValueError`. The decorator owns
+  `_meta["ui"]`; say it with `resource_uri=` and `visibility=`. Other `meta=` keys
+  merge fine alongside.
+
+Neither the TypeScript ext-apps SDK nor FastMCP catches any of these today; we'd
+rather you find out before a host does.
+
+## Beyond inline HTML
+
+`add_html_resource` covers the common case: a string of HTML. For anything else,
+HTML on disk or generated content, build the resource yourself and hand it over:
+
+```python title="server.py" hl_lines="12 18"
+--8<-- "docs_src/apps/tutorial003.py"
+```
+
+`add_resource` fills in the `text/html;profile=mcp-app` MIME type when the resource
+doesn't set one explicitly, and rejects an explicit mismatch: a `ui://` resource
+under any other MIME type is one no host will render.
+
+!!! tip
+    Targeting a pre-GA host that still reads the deprecated flat
+    `_meta["ui/resourceUri"]` key? Merge it yourself:
+    `@apps.tool(resource_uri="ui://x", meta={"ui/resourceUri": "ui://x"})`.
+    The nested `ui` object is the spec shape; the flat key is on its way out.
+
+## See it run
+
+The `apps` story in `examples/stories/` is this page as a runnable pair: a server
+with a UI-bound clock tool and a client that negotiates Apps, reads the tool's
+`_meta.ui.resourceUri`, fetches the HTML, and calls the tool.
+
+```bash
+uv run python -m stories.apps.client
+```

docs/advanced/extensions.md

@@ -0,0 +1,159 @@
+# Extensions
+
+An **extension** is an opt-in bundle of MCP behaviour behind one identifier.
+
+It can contribute tools, resources, and new request methods, and it can wrap `tools/call`.
+The server advertises it under `capabilities.extensions`, the client opts in the same way,
+and nothing changes for anyone who didn't ask for it. That is the contract (SEP-2133), and
+it has one golden rule: **extensions are off by default**.
+
+## Using an extension
+
+Pass instances at construction:
+
+```python title="server.py"
+--8<-- "docs_src/extensions/tutorial001.py"
+```
+
+Done. The server now advertises `io.modelcontextprotocol/ui` under
+`capabilities.extensions` and serves everything the extension contributes.
+
+`Apps` is the built-in reference extension, and it gets its own page: **[MCP Apps](apps.md)**.
+
+!!! note
+    Extensions are fixed at construction. There is no `add_extension` to call later:
+    a server's capability map should not change while clients are connected to it.
+
+The capability map rides `server/discover`, which is a **2026-07-28** path. A legacy
+`initialize` handshake has nowhere to put it, so a legacy client simply doesn't see
+the extension. Design for that: an extension *augments* a server, it must not be the
+only way the server is usable.
+
+## Writing your own
+
+Subclass `Extension` and override only what you need. Every method has a default.
+
+### The identifier
+
+```python
+--8<-- "docs_src/extensions/tutorial002.py"
+```
+
+The identifier is a `vendor-prefix/name` string following the spec's `_meta` key
+grammar: dot-separated labels (each starts with a letter, ends with a letter or
+digit), a slash, then the name. It is validated **when the class is defined**, so a
+typo doesn't wait for a server to boot:
+
+```text
+TypeError: Stamps.identifier must be a `vendor-prefix/name` string
+(reverse-DNS prefix required), got 'stamps'
+```
+
+Use a domain you control as the prefix. `io.modelcontextprotocol/*` is for extensions
+specified by the MCP project itself.
+
+### Contributing tools
+
+The smallest useful extension is one tool and a settings map:
+
+```python title="server.py" hl_lines="17 19-20 22-23 26"
+--8<-- "docs_src/extensions/tutorial003.py"
+```
+
+* `tools()` returns `ToolBinding`s. The server registers each one exactly as if you
+  had called `mcp.add_tool(...)` yourself: same schema generation, same `Context`
+  injection, same everything.
+* `settings()` is the value advertised at `capabilities.extensions["com.example/stamps"]`.
+  Return `{}` (the default) to advertise the extension with no settings.
+* The extension never receives the server. It declares contributions as data;
+  `MCPServer` consumes them. There is no `self.server` to mutate.
+
+And `main()` is the proof, an in-memory client straight against `mcp`:
+
+```python title="server.py" hl_lines="29-34"
+--8<-- "docs_src/extensions/tutorial003.py"
+```
+
+### Serving your own methods
+
+An extension can register **new request methods**: its own verbs, served next to the
+spec's:
+
+```python title="server.py" hl_lines="15-21 30 39-47"
+--8<-- "docs_src/extensions/tutorial004.py"
+```
+
+* `SearchParams` subclasses `RequestParams`, so the 2026 `_meta` envelope parses
+  uniformly and your handler gets validated params, never a raw dict. Bound what
+  the client controls: `Field(ge=1, le=100)` rejects an absurd `limit` before
+  your code allocates anything for it.
+* `require_client_extension(ctx, EXTENSION_ID)` is the gate: a client that did not
+  declare the extension gets the `-32021` (missing required client capability) error,
+  with the machine-readable `requiredCapabilities` payload the spec asks for.
+* `protocol_versions=frozenset({"2026-07-28"})` pins the method to one wire version.
+  At any other version the client gets `METHOD_NOT_FOUND`, exactly as if the method
+  didn't exist there. For that client, it doesn't.
+
+Methods are **strictly additive**. The SDK enforces this at construction, not at
+runtime:
+
+* A `MethodBinding` for a spec-defined method (`tools/list`, `completion/complete`, ...)
+  raises `ValueError` when the binding is constructed. Core verbs belong to the server.
+* Two extensions binding the same method raise when the second one registers.
+  Last-write-wins is how plugins corrupt each other; we don't do that.
+* An empty `protocol_versions` set raises too: a method that can never be served
+  is a bug, not a configuration.
+
+### The client side
+
+The same file's `main()` is the whole client story, both halves of it:
+
+```python title="server.py" hl_lines="53-57"
+--8<-- "docs_src/extensions/tutorial004.py"
+```
+
+* `Client(..., extensions={EXTENSION_ID: {}})` declares the extension. That map
+  becomes `ClientCapabilities.extensions`: on a 2026-07-28 connection it travels in
+  the per-request `_meta` envelope, so the server sees it on **every** request; on
+  a legacy connection it rides the `initialize` handshake. Server code doesn't care
+  which: `require_client_extension(ctx, ...)` and
+  `ctx.session.check_client_capability(...)` read the right source on both paths.
+* Vendor methods drop one layer to `client.session.send_request(...)`; `Client`
+  only grows first-class methods for spec verbs. The `cast` is there because
+  `send_request` is typed against the spec's closed request union.
+
+### Intercepting `tools/call`
+
+The one interceptive hook. Override `intercept_tool_call` to observe, short-circuit,
+or veto a tool call:
+
+```python title="server.py" hl_lines="18-25"
+--8<-- "docs_src/extensions/tutorial005.py"
+```
+
+* `params` is the validated `CallToolRequestParams`: you get `params.name` and
+  `params.arguments` without touching raw JSON.
+* `call_next(ctx)` runs the rest of the chain. Return its result unchanged (observe),
+  return something else (replace), or raise an `MCPError` (refuse).
+* With several extensions, interceptors nest in registration order: the first
+  extension in `extensions=[...]` is outermost.
+* The default implementation is a pass-through, and a server whose extensions never
+  override this hook installs **no** middleware at all. You don't pay for what
+  you don't use.
+
+The hook wraps `tools/call` and nothing else. For every-message concerns, use
+[Middleware](middleware.md). That is what it is for.
+
+## What an extension cannot do
+
+The contribution surface is **closed** on purpose: settings, tools, resources,
+methods, one `tools/call` interceptor. An extension cannot:
+
+* **Reach into the server.** It declares data; it holds no server reference.
+* **Replace core behaviour.** Spec methods are rejected at construction, and
+  `initialize` is reserved by the runner outright.
+* **Register late.** After `MCPServer(...)` returns, the extension set is what it is.
+
+If you are fighting these walls, you are not writing an extension. You are writing
+a fork. The walls are the feature: a user reading `extensions=[Apps(), Stamps()]`
+knows *everything* those two can have touched.

docs/migration.md

@@ -407,6 +407,50 @@ On `ClientSession`, `call_tool` / `get_prompt` / `read_resource` still return th
 
 For protocol 2026-07-28 over Streamable HTTP, a tool's input-schema property may carry an `x-mcp-header` annotation. When a tool the client has listed is called, each annotated argument is mirrored into an `Mcp-Param-<name>` request header (string verbatim, integer as decimal, boolean as `true`/`false`, base64-sentinel-wrapped when not header-safe; `null`/absent arguments are omitted). The argument is also left in the request body. `list_tools` caches a tool's annotations, so list a tool before calling it to enable mirroring; a tool the client never listed emits no `Mcp-Param-*` headers. Other transports ignore the annotation.
 
+### Server extensions API (SEP-2133)
+
+`MCPServer` now accepts opt-in extensions that bundle MCP behaviour behind a
+reverse-DNS identifier and advertise it under `ServerCapabilities.extensions`
+(the 2026-07-28 capability map). An extension subclasses `mcp.server.extension.Extension`
+and overrides only the contribution methods it needs: `tools()`/`resources()`/`methods()`
+(additive) and `intercept_tool_call()` (wraps `tools/call`). The `identifier` must be a
+`vendor-prefix/name` string following the spec's `_meta` key grammar; a class-level
+`identifier` is validated when the subclass is defined, one assigned in `__init__` when
+the extension is registered. Pass instances at construction:
+
+```python
+from mcp.server.mcpserver import MCPServer
+from mcp.server.apps import Apps
+
+mcp = MCPServer("demo", extensions=[Apps()])
+```
+
+The reference extension is `mcp.server.apps.Apps` (`io.modelcontextprotocol/ui`):
+it binds a tool to a `ui://` UI resource via `_meta.ui.resourceUri`, and
+`client_supports_apps(ctx)` gates the SEP-2133 text-only fallback — `True` only
+when the client's ui-extension settings list the `text/html;profile=mcp-app`
+MIME type, per the Apps spec's required `mimeTypes` field. Every
+`@apps.tool(resource_uri=...)` must have a matching resource registered on the
+same `Apps` instance (`add_html_resource` for inline HTML, `add_resource` for a
+pre-built `Resource`); a tool bound to an unregistered URI raises at
+`MCPServer(...)` construction rather than 404ing on `resources/read` at runtime.
+
+Extension methods are strictly additive: a `MethodBinding` cannot name a
+spec-defined request method, and registering one whose method collides with
+another handler raises at construction. A `MethodBinding` may set
+`protocol_versions` to scope an extension method to specific wire versions
+(`frozenset()` is rejected — use `None` to admit every version); a request at
+any other version is `METHOD_NOT_FOUND`. An
+extension handler can call `mcp.server.mcpserver.require_client_extension(ctx, identifier)`
+to reject a request with the `-32021` (missing required client capability) error
+when the client did not declare the extension.
+
+Clients advertise extension support with the new `Client(extensions=...)` /
+`ClientSession(extensions=...)` argument, mirrored into `ClientCapabilities.extensions`.
+The extensions capability map is negotiated over `server/discover` (modern path);
+a legacy `initialize` handshake does not carry it. Extensions are off by default
+and never alter behaviour unless registered.
+
 ### `McpError` renamed to `MCPError`
 
 The `McpError` exception class has been renamed to `MCPError` for consistent naming with the MCP acronym style used throughout the SDK.

docs_src/apps/__init__.py

@@ -0,0 +1,3 @@
+<!doctype html>
+<title>Report</title>
+<p>Quarterly numbers render here.</p>