AWS Bedrock chat model integration.

Setup: Install @langchain/community and set the following environment variables:

npm install @langchain/openai
export BEDROCK_AWS_REGION="your-aws-region"
export BEDROCK_AWS_SECRET_ACCESS_KEY="your-aws-secret-access-key"
export BEDROCK_AWS_ACCESS_KEY_ID="your-aws-access-key-id"

Runtime args can be passed as the second argument to any of the base runnable methods .invoke. .stream, .batch, etc. They can also be passed via .bind, or the second arg in .bindTools, like shown in the examples below:

// When calling `.bind`, call options should be passed via the first argument
const llmWithArgsBound = llm.bind({
stop: ["\n"],
tools: [...],
});

// When calling `.bindTools`, call options should be passed via the second argument
const llmWithTools = llm.bindTools(
[...],
{
stop: ["stop on this token!"],
}
);
Instantiate
import { BedrockChat } from '@langchain/community/chat_models/bedrock';

const llm = new BedrockChat({
region: process.env.BEDROCK_AWS_REGION,
maxRetries: 0,
credentials: {
secretAccessKey: process.env.BEDROCK_AWS_SECRET_ACCESS_KEY!,
accessKeyId: process.env.BEDROCK_AWS_ACCESS_KEY_ID!,
},
model: "anthropic.claude-3-5-sonnet-20240620-v1:0",
temperature: 0,
maxTokens: undefined,
// other params...
});

Invoking
const messages = [
{
type: "system" as const,
content: "You are a helpful translator. Translate the user sentence to French.",
},
{
type: "human" as const,
content: "I love programming.",
},
];
const result = await llm.invoke(messages);
console.log(result);
AIMessage {
  "content": "Here's the translation to French:\n\nJ'adore la programmation.",
  "additional_kwargs": {
    "id": "msg_bdrk_01HCZHa2mKbMZeTeHjLDd286"
  },
  "response_metadata": {
    "type": "message",
    "role": "assistant",
    "model": "claude-3-5-sonnet-20240620",
    "stop_reason": "end_turn",
    "stop_sequence": null,
    "usage": {
      "input_tokens": 25,
      "output_tokens": 19
    }
  },
  "tool_calls": [],
  "invalid_tool_calls": []
}

Streaming Chunks
for await (const chunk of await llm.stream(messages)) {
console.log(chunk);
}
AIMessageChunk {
  "content": "",
  "additional_kwargs": {
    "id": "msg_bdrk_01RhFuGR9uJ2bj5GbdAma4y6"
  },
  "response_metadata": {
    "type": "message",
    "role": "assistant",
    "model": "claude-3-5-sonnet-20240620",
    "stop_reason": null,
    "stop_sequence": null
  },
}
AIMessageChunk {
  "content": "J",
}
AIMessageChunk {
  "content": "'adore la",
}
AIMessageChunk {
  "content": " programmation.",
}
AIMessageChunk {
  "content": "",
  "additional_kwargs": {
    "stop_reason": "end_turn",
    "stop_sequence": null
  },
}
AIMessageChunk {
  "content": "",
  "response_metadata": {
    "amazon-bedrock-invocationMetrics": {
      "inputTokenCount": 25,
      "outputTokenCount": 11,
      "invocationLatency": 659,
      "firstByteLatency": 506
    }
  },
  "usage_metadata": {
    "input_tokens": 25,
    "output_tokens": 11,
    "total_tokens": 36
  }
}

Aggregate Streamed Chunks
import { AIMessageChunk } from '@langchain/core/messages';
import { concat } from '@langchain/core/utils/stream';

const stream = await llm.stream(messages);
let full: AIMessageChunk | undefined;
for await (const chunk of stream) {
full = !full ? chunk : concat(full, chunk);
}
console.log(full);
AIMessageChunk {
  "content": "J'adore la programmation.",
  "additional_kwargs": {
    "id": "msg_bdrk_017b6PuBybA51P5LZ9K6gZHm",
    "stop_reason": "end_turn",
    "stop_sequence": null
  },
  "response_metadata": {
    "type": "message",
    "role": "assistant",
    "model": "claude-3-5-sonnet-20240620",
    "stop_reason": null,
    "stop_sequence": null,
    "amazon-bedrock-invocationMetrics": {
      "inputTokenCount": 25,
      "outputTokenCount": 11,
      "invocationLatency": 1181,
      "firstByteLatency": 1177
    }
  },
  "usage_metadata": {
    "input_tokens": 25,
    "output_tokens": 11,
    "total_tokens": 36
  }
}

Bind tools
import { z } from 'zod';
import { AIMessage } from '@langchain/core/messages';

const GetWeather = {
name: "GetWeather",
description: "Get the current weather in a given location",
schema: z.object({
location: z.string().describe("The city and state, e.g. San Francisco, CA")
}),
}

const GetPopulation = {
name: "GetPopulation",
description: "Get the current population in a given location",
schema: z.object({
location: z.string().describe("The city and state, e.g. San Francisco, CA")
}),
}

const llmWithTools = llm.bindTools([GetWeather, GetPopulation]);
const aiMsg: AIMessage = await llmWithTools.invoke(
"Which city is hotter today and which is bigger: LA or NY?"
);
console.log(aiMsg.tool_calls);
[
  {
    name: 'GetWeather',
    args: { location: 'Los Angeles, CA' },
    id: 'toolu_bdrk_01R2daqwHR931r4baVNzbe38',
    type: 'tool_call'
  },
  {
    name: 'GetWeather',
    args: { location: 'New York, NY' },
    id: 'toolu_bdrk_01WDadwNc7PGqVZvCN7Dr7eD',
    type: 'tool_call'
  },
  {
    name: 'GetPopulation',
    args: { location: 'Los Angeles, CA' },
    id: 'toolu_bdrk_014b8zLkpAgpxrPfewKinJFc',
    type: 'tool_call'
  },
  {
    name: 'GetPopulation',
    args: { location: 'New York, NY' },
    id: 'toolu_bdrk_01Tt8K2MUP15kNuMDFCLEFKN',
    type: 'tool_call'
  }
]

Structured Output
const Joke = z.object({
setup: z.string().describe("The setup of the joke"),
punchline: z.string().describe("The punchline to the joke"),
rating: z.number().optional().describe("How funny the joke is, from 1 to 10")
}).describe('Joke to tell user.');

const structuredLlm = llm.withStructuredOutput(Joke);
const jokeResult = await structuredLlm.invoke("Tell me a joke about cats");
console.log(jokeResult);
{
  setup: "Why don't cats play poker in the jungle?",
  punchline: 'Too many cheetahs!'
}

Response Metadata
const aiMsgForResponseMetadata = await llm.invoke(messages);
console.log(aiMsgForResponseMetadata.response_metadata);
"response_metadata": {
  "type": "message",
  "role": "assistant",
  "model": "claude-3-5-sonnet-20240620",
  "stop_reason": "end_turn",
  "stop_sequence": null,
  "usage": {
    "input_tokens": 25,
    "output_tokens": 19
  }
}

Hierarchy (view full)

Constructors

Properties

codec: EventStreamCodec = ...
credentials: CredentialType

AWS Credentials. If no credentials are provided, the default credentials from @aws-sdk/credential-provider-node will be used.

fetchFn: {
    (input: URL | RequestInfo, init?: RequestInit): Promise<Response>;
    (input: RequestInfo, init?: RequestInit): Promise<Response>;
}

A custom fetch function for low-level access to AWS API. Defaults to fetch().

Type declaration

    • (input, init?): Promise<Response>
    • Parameters

      • input: URL | RequestInfo
      • Optionalinit: RequestInit

      Returns Promise<Response>

    • (input, init?): Promise<Response>
    • Parameters

      • input: RequestInfo
      • Optionalinit: RequestInit

      Returns Promise<Response>

guardrailIdentifier: string = ""

Identifier for the guardrail configuration.

guardrailVersion: string = ""

Version for the guardrail configuration.

model: string = "amazon.titan-tg1-large"

Model to use. For example, "amazon.titan-tg1-large", this is equivalent to the modelId property in the list-foundation-models api.

region: string

The AWS region e.g. us-west-2. Fallback to AWS_DEFAULT_REGION env variable or region specified in ~/.aws/config in case it is not provided here.

streaming: boolean = false

Whether or not to stream responses

usesMessagesApi: boolean = false
endpointHost?: string

Override the default endpoint hostname.

guardrailConfig?: {
    streamProcessingMode: "SYNCHRONOUS" | "ASYNCHRONOUS";
    tagSuffix: string;
}

Required when Guardrail is in use.

maxTokens?: number = undefined

Max tokens.

modelKwargs?: Record<string, unknown>

Additional kwargs to pass to the model.

stopSequences?: string[]

Use as a call option using .bind() instead.

temperature?: number = undefined

Temperature.

trace?: "ENABLED" | "DISABLED"

Trace settings for the Bedrock Guardrails.

Methods

  • Parameters

    • Optionaloptions: unknown

    Returns {
        guardrailConfig: undefined | {
            streamProcessingMode: "SYNCHRONOUS" | "ASYNCHRONOUS";
            tagSuffix: string;
        };
        max_tokens: undefined | number;
        modelKwargs: undefined | Record<string, unknown>;
        stop: any;
        temperature: undefined | number;
        tools: AnthropicTool[];
    }

    • guardrailConfig: undefined | {
          streamProcessingMode: "SYNCHRONOUS" | "ASYNCHRONOUS";
          tagSuffix: string;
      }
    • max_tokens: undefined | number
    • modelKwargs: undefined | Record<string, unknown>
    • stop: any
    • temperature: undefined | number
    • tools: AnthropicTool[]