@crux/anthropic

Peer dependency: @anthropic-ai/sdk

import { createAnthropic, anthropicProviderRuntime, anthropicTranscript, toMessages, fromMessages } from '@crux/anthropic'
import { createGenerateObjectFn, createGenerateTextFn } from '@crux/anthropic'

For usage examples and provider comparison, see the Execution guide.

`createAnthropic(client)`

Create an adapter bound to an Anthropic client.

createAnthropic() is anthropicProviderRuntime.create. The package-owned anthropicTranscript owns tool_use / tool_result conversion plus assistant text/tool-call extraction. Anthropic-specific request params, cache-control system blocks, and response metadata normalization stay owned by @crux/anthropic; Crux still owns prompt resolution, tool loops, safety, validation retry, memory capture, and observability.

Field	Type	Description
`client`	`Anthropic`	The Anthropic client instance

Returns:

Method	Description
`.generate(prompt, options)`	Execute a prompt via `messages.parse` (structured) or `.create` (text)
`.stream(prompt, options)`	Stream a prompt execution

import { createAnthropic } from '@crux/anthropic'
import Anthropic from '@anthropic-ai/sdk'

const adapter = createAnthropic(new Anthropic({ apiKey: process.env.ANTHROPIC_API_KEY }))

const result = await adapter.generate(editDraft, {
  model: 'claude-sonnet-4-5-20250929',
  input: { instruction: 'Fix the intro' },
})
result.text
result._meta.usage

`adapter.generate(prompt, options)`

Field	Type	Description
`prompt`	`Prompt`	The prompt to execute
`options.model`	`string \| FallbackModel<string>`	Model name or `fallback()` wrapper
`options.input`	`object`	Input values
`options.tools`	`Record<string, unknown>?`	Additional Crux tools to merge at call time
`options.toolMiddleware`	`ToolMiddleware \| readonly ToolMiddleware[]?`	Tool execution hooks, including Crux resumable approvals through `result.messages`.
`options.extra.tools`	`ToolUnion[]?`	Anthropic-native tools that bypass Crux tool conversion
`options.extra.tool_choice`	`ToolChoice?`	Anthropic-native tool selection

Returns: normalized Crux generate result with:

result.text — extracted assistant text
result.raw — raw Anthropic SDK response
result._meta — normalized usage, finish reason, tool calls, and model metadata

When the prompt has an output schema, Anthropic's parsed value remains available on result.raw.parsed_output.

Anthropic requires max_tokens. If not set in your prompt's settings, the adapter defaults to 4096.

`adapter.stream(prompt, options)`

Stream a prompt execution.

Field	Type	Description
`prompt`	`Prompt`	The prompt to execute
`options.model`	`string \| FallbackModel<string>`	Model name or `fallback()` wrapper
`options.input`	`object`	Input values
`options.tools`	`Record<string, unknown>?`	Additional Crux tools to merge at call time
`options.toolMiddleware`	`ToolMiddleware \| readonly ToolMiddleware[]?`	Tool execution hooks, including Crux resumable approvals through `result.messages`.
`options.extra.tools`	`ToolUnion[]?`	Anthropic-native tools that bypass Crux tool conversion
`options.extra.tool_choice`	`ToolChoice?`	Anthropic-native tool selection

Returns: Anthropic MessageStream (async iterable of MessageStreamEvent).

`toMessages(sdkMessages)`

Convert Anthropic MessageParam[] to canonical Message[].

Field	Type	Description
`sdkMessages`	`MessageParam[]`	Anthropic SDK messages

Returns: Message[]

`fromMessages(messages)`

Convert canonical Message[] to Anthropic MessageParam[].

Field	Type	Description
`messages`	`Message[]`	Canonical messages

Returns: MessageParam[]

Anthropic has no tool role. Tool results are converted to user messages with tool_result content blocks.

anthropicTranscript is the lower-level NativeTranscriptCodec used by createAnthropic(). toMessages() and fromMessages() are provider-history converters, not a generic cross-provider transcript API; they delegate to the same Anthropic-owned codec that request assembly, assistant tool_use extraction, and second-call tool-loop payloads use.

Assistant tool calls become ordered tool_use content blocks. Tool-result content keeps native Anthropic image and PDF blocks where supported, falling back to deterministic text references for unsupported media.

`createGenerateObjectFn(client, model)`

Create a GenerateObjectFn bound to a client and model.

Field	Type	Description
`client`	`Anthropic`	Anthropic client instance
`model`	`string`	Model name

Returns: GenerateObjectFn

This is a provider-native helper generated from the same native chat profile as createAnthropic(). It uses Anthropic's structured parse surface and returns the parsed { object }, preserving provider errors. It does not run Crux prompt resolution, validation retry, safety, cassettes, tools, memory capture, or instrumentation. Use createGenerateObjectFnFromGenerate(generate) from @crux/core/compaction when a GenerateObjectFn needs full adapter runtime behavior.

`createGenerateTextFn(client, model)`

Create a GenerateTextFn bound to a client and model.

Field	Type	Description
`client`	`Anthropic`	Anthropic client instance
`model`	`string`	Model name

Returns: GenerateTextFn

This helper is also generated from the native chat profile, so text helper calls and adapter calls share request construction and response extraction.

Embeddings and rerankers

@crux/anthropic is a generation adapter. It does not export a provider-native embedding() helper because Anthropic's direct SDK surface here does not provide the dense embedding API that Crux needs for embedding().

When you want Anthropic for generation plus retrieval/indexing around it, use one of these patterns:

pair createAnthropic() with embedding() from @crux/ai
pair it with embedding() from @crux/openai or @crux/google for the write/query vector side
keep Anthropic as the chat model while another provider handles embeddings or reranking

For model-based reranking, use reranker() from @crux/ai.

Types

import type { AnthropicExtra, AnthropicRequest } from '@crux/anthropic'

Guide: Execution
Reference: Prompts
Reference: @crux/ai (Vercel AI SDK)

@crux/anthropic

On this page