Class AnthropicModel

Implements the Anthropic language model integration for Theia. Reasoning-level translation lives in anthropicReasoningFor.

Implements

LanguageModel

Constructors

constructor

new AnthropicModel(
    id: string,
    model: string,
    status: LanguageModelStatus,
    enableStreaming: boolean,
    useCaching: boolean,
    apiKey: () => undefined | string,
    url: undefined | string,
    maxTokens?: number,
    maxRetries?: number,
    proxy?: string,
    reasoningSupport?: ReasoningSupport,
    reasoningApi?: ReasoningApi,
    supportsXHighEffort?: boolean,
    maxInputTokens?: number,
    serverTools?: ServerToolDescriptor[],
    serverSideCompactionSupport?: boolean,
    serverSideCompactionEnabledByDefault?: boolean,
    serverSideCompactionTokenThresholdByDefault?: number,
): AnthropicModel
Parameters
- id: string
- model: string
- status: LanguageModelStatus
- enableStreaming: boolean
- useCaching: boolean
- apiKey: () => undefined | string
- url: undefined | string
- maxTokens: number = DEFAULT_MAX_TOKENS
- maxRetries: number = 3
- Optionalproxy: string
- OptionalreasoningSupport: ReasoningSupport
- OptionalreasoningApi: ReasoningApi
- OptionalsupportsXHighEffort: boolean
- OptionalmaxInputTokens: number
- OptionalserverTools: ServerToolDescriptor[]
- serverSideCompactionSupport: boolean = false
- serverSideCompactionEnabledByDefault: boolean = false
- OptionalserverSideCompactionTokenThresholdByDefault: number
Returns AnthropicModel
- Defined in node/anthropic-language-model.ts:317

Properties

apiKey

apiKey: () => undefined | string

enableStreaming

enableStreaming: boolean

`Readonly`id

id: string

`Optional`maxInputTokens

maxInputTokens?: number

maxRetries

maxRetries: number = 3

maxTokens

maxTokens: number = DEFAULT_MAX_TOKENS

model

model: string

`Optional`proxy

proxy?: string

`Optional`reasoningApi

reasoningApi?: ReasoningApi

`Optional`reasoningSupport

reasoningSupport?: ReasoningSupport

serverSideCompactionEnabledByDefault

serverSideCompactionEnabledByDefault: boolean = false

serverSideCompactionSupport

serverSideCompactionSupport: boolean = false

Whether this model supports provider-native server-side compaction (capability, distinct from whether it is activated).

`Optional`serverSideCompactionTokenThresholdByDefault

serverSideCompactionTokenThresholdByDefault?: number

`Optional`serverTools

serverTools?: ServerToolDescriptor[]

Server tools this model offers, declared code-level by the provider package. Note: If you provide these, you must also provide vendor because server tools are vendor-specific.

status

status: LanguageModelStatus

`Optional`supportsXHighEffort

supportsXHighEffort?: boolean

url

url: undefined | string

useCaching

useCaching: boolean

`Readonly`vendor

vendor: "anthropic" = 'anthropic'

Provider identifier, used to key per-provider settings (e.g. server tool selections) and the capabilities UI.

Methods

`Protected`applyCompactionParams

applyCompactionParams<T extends MessageCreateParams>(
params: T,
request: LanguageModelRequest,
): T
Augments the base message-create params with the beta flag and context-management edit when server-side compaction is enabled for the given request. When disabled, the params are returned unchanged so the default (stable) path is byte-for-byte identical.
Type Parameters
- T extends MessageCreateParams
Parameters
- params: T
- request: LanguageModelRequest
Returns T
- Defined in node/anthropic-language-model.ts:354

`Protected`createServerTools

createServerTools(request: LanguageModelRequest): ToolUnion[]
Translates the enabled server tool ids on the request into native Anthropic server tool params.
Parameters
- request: LanguageModelRequest
Returns ToolUnion[]
- Defined in node/anthropic-language-model.ts:655

`Protected`createTools

createTools(request: LanguageModelRequest): undefined | ToolUnion[]
Parameters
- request: LanguageModelRequest
Returns undefined | ToolUnion[]
- Defined in node/anthropic-language-model.ts:632

`Protected`getSettings

getSettings(request: LanguageModelRequest): Readonly<Record<string, unknown>>
Parameters
- request: LanguageModelRequest
Returns Readonly<Record<string, unknown>>
- Defined in node/anthropic-language-model.ts:338

`Protected`handleNonStreamingRequest

handleNonStreamingRequest(
anthropic: Anthropic,
request: UserRequest,
): Promise<LanguageModelTextResponse>
Parameters
- anthropic: Anthropic
- request: UserRequest
Returns Promise<LanguageModelTextResponse>
- Defined in node/anthropic-language-model.ts:667

`Protected`handleStreamingRequest

handleStreamingRequest(
    anthropic: Anthropic,
    request: UserRequest,
    cancellationToken?: CancellationToken,
    toolMessages?: readonly MessageParam[],
): Promise<LanguageModelStreamResponse>
Parameters
- anthropic: Anthropic
- request: UserRequest
- OptionalcancellationToken: CancellationToken
- OptionaltoolMessages: readonly MessageParam[]
Returns Promise<LanguageModelStreamResponse>
- Defined in node/anthropic-language-model.ts:388

`Protected`initializeAnthropic

initializeAnthropic(): Anthropic
Returns Anthropic
- Defined in node/anthropic-language-model.ts:707

request

request(
request: UserRequest,
cancellationToken?: CancellationToken,
): Promise<LanguageModelResponse>
Parameters
- request: UserRequest
- OptionalcancellationToken: CancellationToken
Returns Promise<LanguageModelResponse>
Implementation of LanguageModel.request
- Defined in node/anthropic-language-model.ts:370

`Protected`useServerSideCompaction

useServerSideCompaction(request: LanguageModelRequest): boolean
Resolves whether this request will route through the Anthropic Beta Messages API for server-side compaction.
Parameters
- request: LanguageModelRequest
Returns boolean
- Defined in node/anthropic-language-model.ts:346

Class AnthropicModel

Implements

Index

Constructors

Properties

Methods

Constructors

constructor

Parameters

Returns AnthropicModel

Properties

apiKey

enableStreaming

Readonlyid

OptionalmaxInputTokens

maxRetries

maxTokens

model

Optionalproxy

OptionalreasoningApi

OptionalreasoningSupport

serverSideCompactionEnabledByDefault

serverSideCompactionSupport

OptionalserverSideCompactionTokenThresholdByDefault

OptionalserverTools

status

OptionalsupportsXHighEffort

url

useCaching

Readonlyvendor

Methods

ProtectedapplyCompactionParams

Type Parameters

Parameters

Returns T

ProtectedcreateServerTools

Parameters

Returns ToolUnion[]

ProtectedcreateTools

Parameters

Returns undefined | ToolUnion[]

ProtectedgetSettings

Parameters

Returns Readonly<Record<string, unknown>>

ProtectedhandleNonStreamingRequest

Parameters

Returns Promise<LanguageModelTextResponse>

ProtectedhandleStreamingRequest

Parameters

Returns Promise<LanguageModelStreamResponse>

ProtectedinitializeAnthropic

Returns Anthropic

request

Parameters

Returns Promise<LanguageModelResponse>

ProtecteduseServerSideCompaction

Parameters

Returns boolean

Settings

On This Page

`Readonly`id

`Optional`maxInputTokens

`Optional`proxy

`Optional`reasoningApi

`Optional`reasoningSupport

`Optional`serverSideCompactionTokenThresholdByDefault

`Optional`serverTools

`Optional`supportsXHighEffort

`Readonly`vendor

`Protected`applyCompactionParams

`Protected`createServerTools

`Protected`createTools

`Protected`getSettings

`Protected`handleNonStreamingRequest

`Protected`handleStreamingRequest

`Protected`initializeAnthropic

`Protected`useServerSideCompaction