Non-streaming chat completions

client.chat.complete() sends one request and waits for the whole response. temperature and maxTokens default to 0.7 and 1000.

Send a completion


import { ServiceProvider } from '@preternatural-js/client-sdk'
 
const result = await client.chat.complete({
  messages: [
    {
      role: 'system',
      content: 'Answer in one paragraph.'
    },
    {
      role: 'user',
      content: 'Why does Saturn have rings?'
    }
  ],
  scope: {
    provider: ServiceProvider.openai,
    modelIdentifier: 'gpt-4o-mini',
    productId: 'your-product-id'
  },
  temperature: 0.7,
  maxTokens: 500
})

Read the response

The result is an OpenAI-compatible ChatCompletion. The text is in choices[0].message.content:


const answer = result.choices[0]?.message.content
 
if (answer) {
  renderAnswer(answer)
}

Handle customer access errors

When the customer is revoked, deleted, or suspended, complete() rejects with CustomerAccessInvalidatedError and the client clears the session:


import {
  CustomerAccessInvalidatedError
} from '@preternatural-js/client-sdk'
 
try {
  const result = await client.chat.complete(request)
  renderAnswer(result.choices[0]?.message.content ?? '')
} catch (error) {
  if (error instanceof CustomerAccessInvalidatedError) {
    redirectToSignIn()
  } else {
    throw error
  }
}