Class: ChatAnthropic

chat_models.ChatAnthropic

Wrapper around Anthropic large language models.

To use you should have the @anthropic-ai/sdk package installed, with the ANTHROPIC_API_KEY environment variable set.

Remarks

Any parameters that are valid to be passed to anthropic.complete can be passed through invocationKwargs, even if not explicitly available on this class.

Hierarchy

• BaseChatModel

  ↳ ChatAnthropic

Implements

• AnthropicInput

Constructors

constructor

• new ChatAnthropic(fields?)

Parameters

Name	Type
fields?	Partial<AnthropicInput> & BaseLanguageModelParams & { anthropicApiKey?: string }

Overrides

BaseChatModel.constructor

Defined in

langchain/src/chat_models/anthropic.ts:130

Properties

apiKey

• Optional apiKey: string

Anthropic API key

Implementation of

AnthropicInput.apiKey

Defined in

langchain/src/chat_models/anthropic.ts:106

---

callbackManager

• callbackManager: CallbackManager

Inherited from

BaseChatModel.callbackManager

Defined in

langchain/src/base_language/index.ts:34

---

caller

• Protected caller: AsyncCaller

The async caller should be used by subclasses to make any async calls, which will thus benefit from the concurrency and retry logic.

Inherited from

BaseChatModel.caller

Defined in

langchain/src/base_language/index.ts:40

---

invocationKwargs

• Optional invocationKwargs: Kwargs

Holds any additional parameters that are valid to pass to anthropic.complete that are not explicitly specified on this class.

Implementation of

AnthropicInput.invocationKwargs

Defined in

langchain/src/chat_models/anthropic.ts:118

---

maxTokensToSample

• maxTokensToSample: number = 2048

A maximum number of tokens to generate before stopping.

Implementation of

AnthropicInput.maxTokensToSample

Defined in

langchain/src/chat_models/anthropic.ts:114

---

modelName

• modelName: string = "claude-v1"

Model name to use

Implementation of

AnthropicInput.modelName

Defined in

langchain/src/chat_models/anthropic.ts:116

---

stopSequences

• Optional stopSequences: string[]

A list of strings upon which to stop generating. You probably want ["\n\nHuman:"], as that's the cue for the next turn in the dialog agent.

Implementation of

AnthropicInput.stopSequences

Defined in

langchain/src/chat_models/anthropic.ts:120

---

streaming

• streaming: boolean = false

Whether to stream the results or not

Implementation of

AnthropicInput.streaming

Defined in

langchain/src/chat_models/anthropic.ts:122

---

temperature

• temperature: number = 1

Amount of randomness injected into the response. Ranges from 0 to 1. Use temp closer to 0 for analytical / multiple choice, and temp closer to 1 for creative and generative tasks.

Implementation of

AnthropicInput.temperature

Defined in

langchain/src/chat_models/anthropic.ts:108

---

topK

• topK: number = -1

Only sample from the top K options for each subsequent token. Used to remove "long tail" low probability responses. Defaults to -1, which disables it.

Implementation of

AnthropicInput.topK

Defined in

langchain/src/chat_models/anthropic.ts:110

---

topP

• topP: number = -1

Does nucleus sampling, in which we compute the cumulative distribution over all the options for each subsequent token in decreasing probability order and cut it off once it reaches a particular probability specified by top_p. Defaults to -1, which disables it. Note that you should either alter temperature or top_p, but not both.

Implementation of

AnthropicInput.topP

Defined in

langchain/src/chat_models/anthropic.ts:112

---

verbose

• verbose: boolean

Whether to print out response text.

Inherited from

BaseChatModel.verbose

Defined in

langchain/src/base_language/index.ts:32

Methods

_combineLLMOutput

▸ _combineLLMOutput(): never[]

Returns

never[]

Overrides

BaseChatModel._combineLLMOutput

Defined in

langchain/src/chat_models/anthropic.ts:296

---

_generate

▸ _generate(messages, stopSequences?): Promise<ChatResult>

Call out to Anthropic's endpoint with k unique prompts

Example

import { ChatAnthropic } from "langchain/llms";
const anthropic = new ChatAnthropic();
const response = await anthropic.generate(new HumanChatMessage(["Tell me a joke."]));

Parameters

Name	Type	Description
messages	BaseChatMessage[]	The messages to pass into the model.
stopSequences?	string[]	Optional list of stop sequences to use when generating.

Returns

Promise<ChatResult>

The full LLM output.

Overrides

BaseChatModel._generate

Defined in

langchain/src/chat_models/anthropic.ts:222

---

_identifyingParams

▸ _identifyingParams(): Object

Get the identifying parameters of the LLM.

Returns

Object

Name	Type
model_name	string

Overrides

BaseChatModel._identifyingParams

Defined in

langchain/src/chat_models/anthropic.ts:177

---

_llmType

▸ _llmType(): string

Returns

string

Overrides

BaseChatModel._llmType

Defined in

langchain/src/chat_models/anthropic.ts:292

---

_modelType

▸ _modelType(): string

Returns

string

Inherited from

BaseChatModel._modelType

Defined in

langchain/src/chat_models/base.ts:76

---

call

▸ call(messages, stop?): Promise<BaseChatMessage>

Parameters

Name	Type
messages	BaseChatMessage[]
stop?	string[]

Returns

Promise<BaseChatMessage>

Inherited from

BaseChatModel.call

Defined in

langchain/src/chat_models/base.ts:97

---

callPrompt

▸ callPrompt(promptValue, stop?): Promise<BaseChatMessage>

Parameters

Name	Type
promptValue	BasePromptValue
stop?	string[]

Returns

Promise<BaseChatMessage>

Inherited from

BaseChatModel.callPrompt

Defined in

langchain/src/chat_models/base.ts:106

---

generate

▸ generate(messages, stop?): Promise<LLMResult>

Parameters

Name	Type
messages	BaseChatMessage[][]
stop?	string[]

Returns

Promise<LLMResult>

Inherited from

BaseChatModel.generate

Defined in

langchain/src/chat_models/base.ts:39

---

generatePrompt

▸ generatePrompt(promptValues, stop?): Promise<LLMResult>

Parameters

Name	Type
promptValues	BasePromptValue[]
stop?	string[]

Returns

Promise<LLMResult>

Inherited from

BaseChatModel.generatePrompt

Defined in

langchain/src/chat_models/base.ts:82

---

getNumTokens

▸ getNumTokens(text): Promise<number>

Parameters

Name	Type
text	string

Returns

Promise<number>

Inherited from

BaseChatModel.getNumTokens

Defined in

langchain/src/base_language/index.ts:62

---

identifyingParams

▸ identifyingParams(): Object

Get the identifying parameters for the model

Returns

Object

Name	Type
model_name	string

Defined in

langchain/src/chat_models/anthropic.ts:187

---

invocationParams

▸ invocationParams(): Omit<SamplingParameters, "prompt"> & Kwargs

Get the parameters used to invoke the model

Returns

Omit<SamplingParameters, "prompt"> & Kwargs

Defined in

langchain/src/chat_models/anthropic.ts:164

---

serialize

▸ serialize(): SerializedLLM

Return a json-like object representing this LLM.

Returns

SerializedLLM

Inherited from

BaseChatModel.serialize

Defined in

langchain/src/base_language/index.ts:92

---

deserialize

▸ Static deserialize(data): Promise<BaseLanguageModel>

Load an LLM from a json-like object describing it.

Parameters

Name	Type
data	SerializedLLM

Returns

Promise<BaseLanguageModel>

Inherited from

BaseChatModel.deserialize

Defined in

langchain/src/base_language/index.ts:103