Errors and requests

ErrorsCopy Anchor Link

Every error from the API is an instance of Axerity.APIError. Specific errors extend it.

Every error has these fields:

Check the specific errors first, then Axerity.APIError.

import Axerity from '@axerity/sdk'

const axerity = new Axerity()

try {
  await axerity.runs.create({ agent: agent.id, input: 'Hello' })
} catch (error) {
  if (error instanceof Axerity.RateLimitError) {
    console.log('retry after', error.retryAfter)
  } else if (error instanceof Axerity.APIError) {
    console.log(error.status, error.code, error.requestId)
  } else {
    throw error
  }
}

RetriesCopy Anchor Link

The SDK retries failed requests with exponential backoff on connection errors, 408, 409, 429, and 5xx responses. Set maxRetries on the client to change how many times it retries.

const axerity = new Axerity({ maxRetries: 3 })

Request optionsCopy Anchor Link

Every method accepts request options as its last argument.

const controller = new AbortController()

const run = await axerity.runs.create(
  { agent: agent.id, input: 'Draft the weekly report' },
  {
    idempotencyKey: 'weekly-report-2026-10-10',
    signal: controller.signal,
    timeout: 120_000,
  },
)

Set timeout on the client to change the default for every request.

IdempotencyCopy Anchor Link

Pass an idempotencyKey when you create something you do not want to create twice, such as a run. If the same request is sent again with the same key, it does not start a second run. Use a key that is unique to the action, such as a job name and a date.

RelatedCopy Anchor Link