AnthropicLLMConfig

Configuration for the Anthropic LLM provider.

Defined in: src/providers/llm/anthropic/AnthropicLLM.ts:79

Configuration for the Anthropic LLM provider.

Remarks

Provide either apiKey (direct API access) or proxyUrl (server-side proxy). At least one must be set; if both are provided, proxyUrl takes precedence.

Anthropic’s API differs from OpenAI’s in that max_tokens is required for every request. This config defaults it to 1024 if not explicitly set.

Example

// Direct API access
const config: AnthropicLLMConfig = {
  apiKey: 'sk-ant-...',
  model: 'claude-haiku-4-5',
  maxTokens: 2048,
  systemPrompt: 'You are a helpful voice assistant.',
};

// Via server-side proxy (recommended for browser apps)
const proxyConfig: AnthropicLLMConfig = {
  proxyUrl: 'http://localhost:3000/api/proxy/anthropic',
  model: 'claude-sonnet-4-6',
};

See

LLMProviderConfig for inherited base properties (temperature, topP, systemPrompt, etc.).

Extends

LLMProviderConfig

Properties

Property	Type	Default value	Description	Overrides	Inherited from	Defined in
`apiKey?`	`string`	`undefined`	API key or authentication token for the provider. Remarks For client-side usage, consider using a proxy server to keep API keys secure. The SDK provides Express, Next.js, and Node adapters for this purpose.	-	`LLMProviderConfig`.`apiKey`	src/core/types/providers.ts:67
`authType?`	`"token"` \| `"bearer"`	`Provider-specific (typically 'token' for Deepgram, ignored for REST providers)`	Authentication type for providers that support multiple auth mechanisms. Remarks Controls how the `apiKey` is sent to the provider: - `'token'` — WebSocket subprotocol `['token', apiKey]` or header `Authorization: Token <key>`. This is the default for Deepgram providers. - `'bearer'` — WebSocket subprotocol `['bearer', token]` or header `Authorization: Bearer <token>`. Use this for OAuth tokens or providers that expect Bearer auth. REST/SDK providers (Anthropic, OpenAI) handle auth through their SDK constructors and ignore this field.	-	`LLMProviderConfig`.`authType`	src/core/types/providers.ts:111
`debug?`	`boolean`	`false`	Whether to enable debug logging for this provider. Remarks When `true`, the provider emits detailed internal logs. This is separate from the SDK-level LoggingConfig.	-	`LLMProviderConfig`.`debug`	src/core/types/providers.ts:122
`endpoint?`	`string`	`undefined`	Custom endpoint URL to override the provider’s default API endpoint. Remarks Useful for self-hosted instances, proxy servers, or development environments.	-	`LLMProviderConfig`.`endpoint`	src/core/types/providers.ts:75
`maxRetries?`	`number`	`3`	Maximum number of retries for failed API requests.	-	-	src/providers/llm/anthropic/AnthropicLLM.ts:106
`maxTokens?`	`number`	`1024`	Maximum tokens to generate per response. Remarks Anthropic’s Messages API requires this field on every request. The provider defaults to `1024` if not set in config or per-call options.	`LLMProviderConfig`.`maxTokens`	-	src/providers/llm/anthropic/AnthropicLLM.ts:100
`model`	`string`	`'claude-haiku-4-5'`	Anthropic model identifier. Remarks - Fastest: `'claude-haiku-4-5'` (default) - Balanced: `'claude-sonnet-4-6'` - Most capable: `'claude-opus-4-6'`	`LLMProviderConfig`.`model`	-	src/providers/llm/anthropic/AnthropicLLM.ts:90
`proxyUrl?`	`string`	`undefined`	URL of a CompositeVoice proxy server endpoint for this provider. Remarks When set, requests are routed through the proxy which injects the real API key server-side. This keeps API keys out of the browser. For WebSocket providers the HTTP URL is automatically converted to `ws(s)://`. At least one of `apiKey` or `proxyUrl` must be set for providers that require authentication (all except NativeSTT, NativeTTS, and WebLLM). Example `proxyUrl: 'http://localhost:3000/api/proxy/deepgram'`	-	`LLMProviderConfig`.`proxyUrl`	src/core/types/providers.ts:93
`stopSequences?`	`string`[]	`undefined`	Sequences that cause the LLM to stop generating. Remarks When the model generates any of these sequences, generation halts. Useful for controlling response boundaries.	-	`LLMProviderConfig`.`stopSequences`	src/core/types/providers.ts:715
`stream?`	`boolean`	`undefined`	Whether to stream the LLM response token by token. Remarks When `true`, the provider yields tokens incrementally via an async iterable. Streaming is essential for low-latency voice applications as it allows TTS to begin synthesizing before the full response is generated.	-	`LLMProviderConfig`.`stream`	src/core/types/providers.ts:706
`systemPrompt?`	`string`	`undefined`	System prompt providing instructions and context to the LLM. Remarks Sets the behavior and persona of the assistant. For voice applications, include instructions to keep responses brief and conversational.	-	`LLMProviderConfig`.`systemPrompt`	src/core/types/providers.ts:696
`temperature?`	`number`	`undefined`	Temperature for controlling generation randomness. Remarks Values from 0 (deterministic) to 2 (highly creative). Lower values produce more focused responses; higher values increase variety.	-	`LLMProviderConfig`.`temperature`	src/core/types/providers.ts:668
`timeout?`	`number`	`undefined`	Request timeout in milliseconds. Remarks Applies to HTTP requests (REST providers) and connection establishment (WebSocket providers). Set to `0` for no timeout.	-	`LLMProviderConfig`.`timeout`	src/core/types/providers.ts:131
`topP?`	`number`	`undefined`	Top-P (nucleus) sampling parameter. Remarks Limits token selection to the smallest set whose cumulative probability exceeds this value. Values from 0 to 1. Often used as an alternative to temperature.	-	`LLMProviderConfig`.`topP`	src/core/types/providers.ts:687