design: add design for rate limiting (#36)

I added an example for how rate limiting should be implemented using
middleware.

Fixes #22 

## Motivation and Context
<!-- Why is this change needed? What problem does it solve? -->
Rate limiting can be enabled using middleware. This adds an example on
how to do that.

## How Has This Been Tested?
<!-- Have you tested this in a real application? Which scenarios were
tested? -->
N/A

## Breaking Changes
<!-- Will users need to update their code or configurations? -->
No

## Types of changes
<!-- What types of changes does your code introduce? Put an `x` in all
the boxes that apply: -->
- [ ] Bug fix (non-breaking change which fixes an issue)
- [ ] New feature (non-breaking change which adds functionality)
- [ ] Breaking change (fix or feature that would cause existing
functionality to change)
- [x] Documentation update

## Checklist
<!-- Go over all the following points, and put an `x` in all the boxes
that apply. -->
- [x] I have read the [MCP
Documentation](https://modelcontextprotocol.io)
- [x] My code follows the repository's style guidelines
- [x] New and existing tests pass locally
- [x] I have added appropriate error handling
- [x] I have added or updated documentation as needed

## Additional context
<!-- Add any other context, implementation notes, or design decisions
-->

design/design.md

@@ -470,6 +470,10 @@ server.AddReceivingMiddleware(withLogging)
 
 **Differences from mcp-go**: Version 0.26.0 of mcp-go defines 24 server hooks. Each hook consists of a field in the `Hooks` struct, a `Hooks.Add` method, and a type for the hook function. These are rarely used. The most common is `OnError`, which occurs fewer than ten times in open-source code.
 
+#### Rate Limiting
+
+Rate limiting can be configured using middleware. Please see [examples/rate-limiting](<https://github.com/modelcontextprotocol/go-sdk/tree/main/examples/rate-limiting>] for an example on how to implement this.
+
 ### Errors
 
 With the exception of tool handler errors, protocol errors are handled transparently as Go errors: errors in server-side feature handlers are propagated as errors from calls from the `ClientSession`, and vice-versa.

examples/rate-limiting/go.mod

@@ -0,0 +1,8 @@
+module github.com/modelcontextprotocol/go-sdk/examples/rate-limiting
+
+go 1.25
+
+require (
+	github.com/modelcontextprotocol/go-sdk v0.0.0-20250625185707-09181c2c2e89
+	golang.org/x/time v0.12.0
+)

examples/rate-limiting/go.sum

@@ -0,0 +1,8 @@
+github.com/google/go-cmp v0.7.0 h1:wk8382ETsv4JYUZwIsn6YpYiWiBsYLSJiTsyBybVuN8=
+github.com/google/go-cmp v0.7.0/go.mod h1:pXiqmnSA92OHEEa9HXL2W4E7lf9JzCmGVUdgjX3N/iU=
+github.com/modelcontextprotocol/go-sdk v0.0.0-20250625185707-09181c2c2e89 h1:kUGBYP25FTv3ZRBhLT4iQvtx4FDl7hPkWe3isYrMxyo=
+github.com/modelcontextprotocol/go-sdk v0.0.0-20250625185707-09181c2c2e89/go.mod h1:DcXfbr7yl7e35oMpzHfKw2nUYRjhIGS2uou/6tdsTB0=
+golang.org/x/time v0.12.0 h1:ScB/8o8olJvc+CQPWrK3fPZNfh7qgwCrY0zJmoEQLSE=
+golang.org/x/time v0.12.0/go.mod h1:CDIdPxbZBQxdj6cxyCIdrNogrJKMJ7pr37NYpMcMDSg=
+golang.org/x/tools v0.34.0 h1:qIpSLOxeCYGg9TrcJokLBG4KFA6d795g0xkBkiESGlo=
+golang.org/x/tools v0.34.0/go.mod h1:pAP9OwEaY1CAW3HOmg3hLZC5Z0CCmzjAF2UQMSqNARg=

examples/rate-limiting/main.go

@@ -0,0 +1,54 @@
+// Copyright 2025 The Go MCP SDK Authors. All rights reserved.
+// Use of this source code is governed by an MIT-style
+// license that can be found in the LICENSE file.
+
+package main
+
+import (
+	"context"
+	"errors"
+	"time"
+
+	"github.com/modelcontextprotocol/go-sdk/mcp"
+	"golang.org/x/time/rate"
+)
+
+// GlobalRateLimiterMiddleware creates a middleware that applies a global rate limit.
+// Every request attempting to pass through will try to acquire a token.
+// If a token cannot be acquired immediately, the request will be rejected.
+func GlobalRateLimiterMiddleware[S mcp.Session](limiter *rate.Limiter) mcp.Middleware[S] {
+	return func(next mcp.MethodHandler[S]) mcp.MethodHandler[S] {
+		return func(ctx context.Context, session S, method string, params mcp.Params) (mcp.Result, error) {
+			if !limiter.Allow() {
+				return nil, errors.New("JSON RPC overloaded")
+			}
+			return next(ctx, session, method, params)
+		}
+	}
+}
+
+// PerMethodRateLimiterMiddleware creates a middleware that applies rate limiting
+// on a per-method basis.
+// Methods not specified in limiters will not be rate limited by this middleware.
+func PerMethodRateLimiterMiddleware[S mcp.Session](limiters map[string]*rate.Limiter) mcp.Middleware[S] {
+	return func(next mcp.MethodHandler[S]) mcp.MethodHandler[S] {
+		return func(ctx context.Context, session S, method string, params mcp.Params) (mcp.Result, error) {
+			if limiter, ok := limiters[method]; ok {
+				if !limiter.Allow() {
+					return nil, errors.New("JSON RPC overloaded")
+				}
+			}
+			return next(ctx, session, method, params)
+		}
+	}
+}
+
+func main() {
+	server := mcp.NewServer("greeter1", "v0.0.1", nil)
+	server.AddReceivingMiddleware(GlobalRateLimiterMiddleware[*mcp.ServerSession](rate.NewLimiter(rate.Every(time.Second/5), 10)))
+	server.AddReceivingMiddleware(PerMethodRateLimiterMiddleware[*mcp.ServerSession](map[string]*rate.Limiter{
+		"callTool":  rate.NewLimiter(rate.Every(time.Second), 5),  // once a second with a burst up to 5
+		"listTools": rate.NewLimiter(rate.Every(time.Minute), 20), // once a minute with a burst up to 20
+	}))
+	// Run Server logic.
+}