Skip to main content
Middleware specifically designed for models hosted on AWS Bedrock. Learn more about middleware.
MiddlewareDescription
Prompt cachingReduce costs by caching repetitive prompt prefixes

​
Prompt caching

Reduce inference latency and input token costs by caching frequently reused prompt prefixes on Amazon Bedrock. bedrockPromptCachingMiddleware enables caching through modelSettings. ChatBedrockConverse then translates that into the correct AWS wire format at request time. Cache checkpoints are placed after the system prompt, tool definitions, and the most recent message where supported, so that the model can skip recomputation of previously seen content on subsequent requests. Cache placement varies by model family: for example, Nova skips some tool definition and tool-result cases. Prompt caching is useful for the following:
  • Multi-turn conversations with long, consistent system prompts
  • Agents with many tool definitions that remain constant across invocations
  • Document-based Q&A where users ask multiple questions over the same uploaded context
  • Batch processing workloads with repeated static content
Supported models:
  • Anthropic Claude
  • Amazon Nova
Learn more about AWS Bedrock prompt caching strategies and limitations. Cached content must exceed 1,024 tokens for a cache checkpoint to take effect, sometimes more depending on model. See supported models, regions, and limits.
API reference: BedrockPromptCachingMiddleware
Model string
import { createAgent, bedrockPromptCachingMiddleware } from "langchain";

const agent = createAgent({
  model: "bedrock:us.anthropic.claude-sonnet-4-5-20250929-v1:0",
  systemPrompt: "<Your long system prompt here>",
  middleware: [bedrockPromptCachingMiddleware({ ttl: "1h" })],
});
ChatBedrockConverse instance
import { createAgent, bedrockPromptCachingMiddleware } from "langchain";
import { ChatBedrockConverse } from "@langchain/aws";

const agent = createAgent({
  model: new ChatBedrockConverse({ model: "us.anthropic.claude-sonnet-4-5-20250929-v1:0" }),
  systemPrompt: "<Your long system prompt here>",
  middleware: [bedrockPromptCachingMiddleware({ ttl: "5m" })],
});
​
enableCaching
boolean
default:"true"
Whether to apply prompt caching.
​
ttl
string
default:"5m"
Time to live for cached content. Valid values: '5m' or '1h'. Note that Amazon Nova models only support '5m'.
​
minMessagesToCache
number
default:"1"
Minimum number of messages before caching starts. The system prompt counts as one message.
​
unsupportedModelBehavior
string
default:"warn"
Behavior when using unsupported models. Options: 'ignore', 'warn', or 'raise'.
The middleware caches content up to and including the latest message in each request. On subsequent requests within the TTL window (5 minutes or 1 hour), previously seen content is retrieved from cache rather than reprocessed, reducing costs and latency.How it works:
  1. First request: System prompt, tools, and the user message are sent to the API and cached
  2. Second request: The cached content is retrieved from cache. Only the new message needs to be processed
  3. This pattern continues for each turn, with each request reusing the cached conversation history
Prompt caching reduces API costs by caching tokens, but does not provide conversation memory. To persist conversation history across invocations, use a checkpointer like MemorySaver.
import { createAgent, bedrockPromptCachingMiddleware, AIMessage, HumanMessage, tool } from "langchain";
import { ChatBedrockConverse } from "@langchain/aws";
import { z } from "zod";

const getWeather = tool(
  async ({ city }) => `The weather in ${city} is sunny and 72F.`,
  {
    name: "get_weather",
    description: "Get the current weather for a city.",
    schema: z.object({ city: z.string() }),
  }
);

// System prompt must exceed 1,024 tokens for caching to take effect
const LONG_PROMPT =
  "You are a helpful weather assistant with deep expertise in meteorology, " +
  "climate science, and atmospheric phenomena. When answering questions about " +
  "weather, provide accurate and up-to-date information. " +
  "You should always strive to give the most helpful response possible. ".repeat(85);

const agent = createAgent({
  model: new ChatBedrockConverse({ model: "us.anthropic.claude-sonnet-4-5-20250929-v1:0" }),
  systemPrompt: LONG_PROMPT,
  tools: [getWeather],
  middleware: [bedrockPromptCachingMiddleware({ ttl: "5m" })],
});

// First invocation: writes the cache (system prompt, tool definitions, and message)
let response = await agent.invoke({
  messages: [new HumanMessage("What is the weather in Miami?")],
});
const last = response.messages.at(-1);
console.log(last?.content);

// Check cache token usage
if (AIMessage.isInstance(last)) {
  const details = last.usage_metadata?.input_token_details;
  if (details) {
    console.log(`Cache read: ${details.cache_read ?? 0}, Cache write: ${details.cache_creation ?? 0}`);
  }
}

// Second invocation within the TTL: reuses the cached system prompt and tool definitions
response = await agent.invoke({
  messages: [new HumanMessage("How about Seattle?")],
});
console.log(response.messages.at(-1)?.content);

​
Model-specific behavior

The middleware handles differences between model families automatically:
FeatureChatBedrockConverse (Anthropic)ChatBedrockConverse (Nova)
System prompt caching✅✅
Tool definition caching✅❌
Message caching✅✅ (excludes tool result messages)
Extended TTL (1h)✅❌
Connect these docs to Claude, VSCode, and more via MCP for real-time answers.
Edit this page on GitHub or file an issue.