Anthropic SDK
Import from
supercompat/anthropic to get an Anthropic instance. It exposes the Messages API — client.messages.create() — and returns objects typed as Anthropic.Message.Install
npm install supercompat @anthropic-ai/sdk
Minimal call
import Anthropic from '@anthropic-ai/sdk'
import {
supercompat,
anthropicClientAdapter,
} from 'supercompat/anthropic'
const client = supercompat({
clientAdapter: anthropicClientAdapter({ anthropic: new Anthropic() }),
})
const message = await client.messages.create({
model: 'claude-sonnet-4-6',
max_tokens: 512,
messages: [
{ role: 'user', content: 'Explain the Messages API in one sentence.' },
],
})
console.log(message.content)
The Messages API is stateless — you pass the full conversation history on each call.
Streaming
const stream = await client.messages.stream({
model: 'claude-sonnet-4-6',
max_tokens: 512,
messages: [{ role: 'user', content: 'Count to five.' }],
})
for await (const event of stream) {
if (event.type === 'content_block_delta' && event.delta.type === 'text_delta') {
process.stdout.write(event.delta.text)
}
}
Against a different backend
The Anthropic SDK surface still works when you point it at another provider through the client adapter. For example, to route
client.messages.create() to OpenAI:import OpenAI from 'openai'
import {
supercompat,
openaiClientAdapter,
} from 'supercompat/anthropic'
const client = supercompat({
clientAdapter: openaiClientAdapter({ openai: new OpenAI() }),
})
const message = await client.messages.create({
model: 'gpt-4.1-mini',
max_tokens: 512,
messages: [{ role: 'user', content: 'Explain OpenAI in one sentence.' }],
})
The return type stays
Anthropic.Message even though the backend is OpenAI.When to use this surface
Pick the Anthropic output SDK when:
Your app is already structured around
client.messages.create() and Anthropic's content blocks.You want to keep Anthropic's streaming event vocabulary throughout your code.
For new projects, the OpenAI output SDK with the Responses API is generally the recommended default.