Merge branch 'main' into main

Author: jingyugao

.gitignore

@@ -166,4 +166,5 @@ cython_debug/
 
 # vscode
 .vscode/
+.windsurfrules
 **/CLAUDE.local.md

CLAUDE.md

@@ -19,7 +19,7 @@ This document contains critical information about working with this codebase. Fo
    - Line length: 88 chars maximum
 
 3. Testing Requirements
-   - Framework: `uv run pytest`
+   - Framework: `uv run --frozen pytest`
    - Async testing: use anyio, not asyncio
    - Coverage: test edge cases and errors
    - New features require tests
@@ -54,9 +54,9 @@ This document contains critical information about working with this codebase. Fo
 ## Code Formatting
 
 1. Ruff
-   - Format: `uv run ruff format .`
-   - Check: `uv run ruff check .`
-   - Fix: `uv run ruff check . --fix`
+   - Format: `uv run --frozen ruff format .`
+   - Check: `uv run --frozen ruff check .`
+   - Fix: `uv run --frozen ruff check . --fix`
    - Critical issues:
      - Line length (88 chars)
      - Import sorting (I001)
@@ -67,7 +67,7 @@ This document contains critical information about working with this codebase. Fo
      - Imports: split into multiple lines
 
 2. Type Checking
-   - Tool: `uv run pyright`
+   - Tool: `uv run --frozen pyright`
    - Requirements:
      - Explicit None checks for Optional
      - Type narrowing for strings
@@ -104,6 +104,10 @@ This document contains critical information about working with this codebase. Fo
      - Add None checks
      - Narrow string types
      - Match existing patterns
+   - Pytest:
+     - If the tests aren't finding the anyio pytest mark, try adding PYTEST_DISABLE_PLUGIN_AUTOLOAD=""
+       to the start of the pytest run command eg:
+       `PYTEST_DISABLE_PLUGIN_AUTOLOAD="" uv run --frozen pytest`
 
 3. Best Practices
    - Check git status before commits

README.md

@@ -309,6 +309,33 @@ async def long_task(files: list[str], ctx: Context) -> str:
     return "Processing complete"
 ```
 
+### Authentication
+
+Authentication can be used by servers that want to expose tools accessing protected resources.
+
+`mcp.server.auth` implements an OAuth 2.0 server interface, which servers can use by
+providing an implementation of the `OAuthServerProvider` protocol.
+
+```
+mcp = FastMCP("My App",
+        auth_provider=MyOAuthServerProvider(),
+        auth=AuthSettings(
+            issuer_url="https://myapp.com",
+            revocation_options=RevocationOptions(
+                enabled=True,
+            ),
+            client_registration_options=ClientRegistrationOptions(
+                enabled=True,
+                valid_scopes=["myscope", "myotherscope"],
+                default_scopes=["myscope"],
+            ),
+            required_scopes=["myscope"],
+        ),
+)
+```
+
+See [OAuthServerProvider](mcp/server/auth/provider.py) for more details.
+
 ## Running Your Server
 
 ### Development Mode

examples/clients/simple-chatbot/mcp_simple_chatbot/main.py

@@ -122,8 +122,10 @@ async def list_tools(self) -> list[Any]:
 
         for item in tools_response:
             if isinstance(item, tuple) and item[0] == "tools":
-                for tool in item[1]:
-                    tools.append(Tool(tool.name, tool.description, tool.inputSchema))
+                tools.extend(
+                    Tool(tool.name, tool.description, tool.inputSchema)
+                    for tool in item[1]
+                )
 
         return tools
 
@@ -282,10 +284,9 @@ def __init__(self, servers: list[Server], llm_client: LLMClient) -> None:
 
     async def cleanup_servers(self) -> None:
         """Clean up all servers properly."""
-        cleanup_tasks = []
-        for server in self.servers:
-            cleanup_tasks.append(asyncio.create_task(server.cleanup()))
-
+        cleanup_tasks = [
+            asyncio.create_task(server.cleanup()) for server in self.servers
+        ]
         if cleanup_tasks:
             try:
                 await asyncio.gather(*cleanup_tasks, return_exceptions=True)
@@ -322,8 +323,7 @@ async def process_llm_response(self, llm_response: str) -> str:
                                 total = result["total"]
                                 percentage = (progress / total) * 100
                                 logging.info(
-                                    f"Progress: {progress}/{total} "
-                                    f"({percentage:.1f}%)"
+                                    f"Progress: {progress}/{total} ({percentage:.1f}%)"
                                 )
 
                             return f"Tool execution result: {result}"

examples/servers/simple-prompt/mcp_simple_prompt/server.py

@@ -90,6 +90,7 @@ async def get_prompt(
     if transport == "sse":
         from mcp.server.sse import SseServerTransport
         from starlette.applications import Starlette
+        from starlette.responses import Response
         from starlette.routing import Mount, Route
 
         sse = SseServerTransport("/messages/")
@@ -101,6 +102,7 @@ async def handle_sse(request):
                 await app.run(
                     streams[0], streams[1], app.create_initialization_options()
                 )
+            return Response()
 
         starlette_app = Starlette(
             debug=True,

examples/servers/simple-resource/mcp_simple_resource/server.py

@@ -46,6 +46,7 @@ async def read_resource(uri: FileUrl) -> str | bytes:
     if transport == "sse":
         from mcp.server.sse import SseServerTransport
         from starlette.applications import Starlette
+        from starlette.responses import Response
         from starlette.routing import Mount, Route
 
         sse = SseServerTransport("/messages/")
@@ -57,11 +58,12 @@ async def handle_sse(request):
                 await app.run(
                     streams[0], streams[1], app.create_initialization_options()
                 )
+            return Response()
 
         starlette_app = Starlette(
             debug=True,
             routes=[
-                Route("/sse", endpoint=handle_sse),
+                Route("/sse", endpoint=handle_sse, methods=["GET"]),
                 Mount("/messages/", app=sse.handle_post_message),
             ],
         )

examples/servers/simple-streamablehttp-stateless/README.md

@@ -0,0 +1,41 @@
+# MCP Simple StreamableHttp Stateless Server Example
+
+A stateless MCP server example demonstrating the StreamableHttp transport without maintaining session state. This example is ideal for understanding how to deploy MCP servers in multi-node environments where requests can be routed to any instance.
+
+## Features
+
+- Uses the StreamableHTTP transport in stateless mode (mcp_session_id=None)
+- Each request creates a new ephemeral connection
+- No session state maintained between requests
+- Task lifecycle scoped to individual requests
+- Suitable for deployment in multi-node environments
+
+
+## Usage
+
+Start the server:
+
+```bash
+# Using default port 3000
+uv run mcp-simple-streamablehttp-stateless
+
+# Using custom port
+uv run mcp-simple-streamablehttp-stateless --port 3000
+
+# Custom logging level
+uv run mcp-simple-streamablehttp-stateless --log-level DEBUG
+
+# Enable JSON responses instead of SSE streams
+uv run mcp-simple-streamablehttp-stateless --json-response
+```
+
+The server exposes a tool named "start-notification-stream" that accepts three arguments:
+
+- `interval`: Time between notifications in seconds (e.g., 1.0)
+- `count`: Number of notifications to send (e.g., 5)
+- `caller`: Identifier string for the caller
+
+
+## Client
+
+You can connect to this server using an HTTP client. For now, only the TypeScript SDK has streamable HTTP client examples, or you can use [Inspector](https://github.com/modelcontextprotocol/inspector) for testing.

examples/servers/simple-streamablehttp-stateless/mcp_simple_streamablehttp_stateless/__init__.py

@@ -0,0 +1,4 @@
+from .server import main
+
+if __name__ == "__main__":
+    main()

examples/servers/simple-streamablehttp-stateless/mcp_simple_streamablehttp_stateless/__main__.py

@@ -0,0 +1,168 @@
+import contextlib
+import logging
+
+import anyio
+import click
+import mcp.types as types
+from mcp.server.lowlevel import Server
+from mcp.server.streamableHttp import (
+    StreamableHTTPServerTransport,
+)
+from starlette.applications import Starlette
+from starlette.routing import Mount
+
+logger = logging.getLogger(__name__)
+# Global task group that will be initialized in the lifespan
+task_group = None
+
+
+@contextlib.asynccontextmanager
+async def lifespan(app):
+    """Application lifespan context manager for managing task group."""
+    global task_group
+
+    async with anyio.create_task_group() as tg:
+        task_group = tg
+        logger.info("Application started, task group initialized!")
+        try:
+            yield
+        finally:
+            logger.info("Application shutting down, cleaning up resources...")
+            if task_group:
+                tg.cancel_scope.cancel()
+                task_group = None
+            logger.info("Resources cleaned up successfully.")
+
+
+@click.command()
+@click.option("--port", default=3000, help="Port to listen on for HTTP")
+@click.option(
+    "--log-level",
+    default="INFO",
+    help="Logging level (DEBUG, INFO, WARNING, ERROR, CRITICAL)",
+)
+@click.option(
+    "--json-response",
+    is_flag=True,
+    default=False,
+    help="Enable JSON responses instead of SSE streams",
+)
+def main(
+    port: int,
+    log_level: str,
+    json_response: bool,
+) -> int:
+    # Configure logging
+    logging.basicConfig(
+        level=getattr(logging, log_level.upper()),
+        format="%(asctime)s - %(name)s - %(levelname)s - %(message)s",
+    )
+
+    app = Server("mcp-streamable-http-stateless-demo")
+
+    @app.call_tool()
+    async def call_tool(
+        name: str, arguments: dict
+    ) -> list[types.TextContent | types.ImageContent | types.EmbeddedResource]:
+        ctx = app.request_context
+        interval = arguments.get("interval", 1.0)
+        count = arguments.get("count", 5)
+        caller = arguments.get("caller", "unknown")
+
+        # Send the specified number of notifications with the given interval
+        for i in range(count):
+            await ctx.session.send_log_message(
+                level="info",
+                data=f"Notification {i+1}/{count} from caller: {caller}",
+                logger="notification_stream",
+                related_request_id=ctx.request_id,
+            )
+            if i < count - 1:  # Don't wait after the last notification
+                await anyio.sleep(interval)
+
+        return [
+            types.TextContent(
+                type="text",
+                text=(
+                    f"Sent {count} notifications with {interval}s interval"
+                    f" for caller: {caller}"
+                ),
+            )
+        ]
+
+    @app.list_tools()
+    async def list_tools() -> list[types.Tool]:
+        return [
+            types.Tool(
+                name="start-notification-stream",
+                description=(
+                    "Sends a stream of notifications with configurable count"
+                    " and interval"
+                ),
+                inputSchema={
+                    "type": "object",
+                    "required": ["interval", "count", "caller"],
+                    "properties": {
+                        "interval": {
+                            "type": "number",
+                            "description": "Interval between notifications in seconds",
+                        },
+                        "count": {
+                            "type": "number",
+                            "description": "Number of notifications to send",
+                        },
+                        "caller": {
+                            "type": "string",
+                            "description": (
+                                "Identifier of the caller to include in notifications"
+                            ),
+                        },
+                    },
+                },
+            )
+        ]
+
+    # ASGI handler for stateless HTTP connections
+    async def handle_streamable_http(scope, receive, send):
+        logger.debug("Creating new transport")
+        # Use lock to prevent race conditions when creating new sessions
+        http_transport = StreamableHTTPServerTransport(
+            mcp_session_id=None,
+            is_json_response_enabled=json_response,
+        )
+        async with http_transport.connect() as streams:
+            read_stream, write_stream = streams
+
+            if not task_group:
+                raise RuntimeError("Task group is not initialized")
+
+            async def run_server():
+                await app.run(
+                    read_stream,
+                    write_stream,
+                    app.create_initialization_options(),
+                    # Runs in standalone mode for stateless deployments
+                    # where clients perform initialization with any node
+                    standalone_mode=True,
+                )
+
+            # Start server task
+            task_group.start_soon(run_server)
+
+            # Handle the HTTP request and return the response
+            await http_transport.handle_request(scope, receive, send)
+
+    # Create an ASGI application using the transport
+    starlette_app = Starlette(
+        debug=True,
+        routes=[
+            Mount("/mcp", app=handle_streamable_http),
+        ],
+        lifespan=lifespan,
+    )
+
+    import uvicorn
+
+    uvicorn.run(starlette_app, host="0.0.0.0", port=port)
+
+    return 0