generated: '2026-07-18'
method: searched
source: https://docs.kits.ai/api-reference
description: >-
  Cross-cutting request/response conventions for the Kits AI API, captured from
  the published documentation and derived from the OpenAPI. The API is an
  asynchronous job model: POST create endpoints enqueue a job and return its
  status; clients poll the matching GET endpoint until status is "success" and
  signed output file URLs are populated.
authentication:
  style: bearer
  header: Authorization
  format: 'Bearer <api-key>'
  key_management: https://app.kits.ai/api-access
  ref: authentication/arpeggi-labs-authentication.yml
async_jobs:
  pattern: create-then-poll
  statuses: [running, success, error, cancelled]
  output: >-
    Signed output file URLs (outputFileUrl, vocalAudioFileUrl, stemFileUrls,
    etc.) are populated once status is "success"; signed URLs expire in 4 hours.
uploads:
  create_content_type: multipart/form-data
  note: >-
    Voice conversion, stem splitter, and vocal separation create endpoints
    require multipart form uploads; the voice blender create endpoint accepts
    application/json.
  file_formats: [wav, mp3, flac, webm]
  max_file_size:
    voice_conversion: 100MB
    stem_splitter: 50MB
    vocal_separation: 50MB
pagination:
  style: page-number
  params:
    page: Page offset (default 1)
    perPage: Page size (default 10)
    order: asc (default) or desc
  response_envelope:
    data: array of resources
    meta:
      fields: [total, perPage, currentPage, lastPage, firstPage, firstPageUrl, lastPageUrl, nextPageUrl, previousPageUrl]
  ref: data-model/arpeggi-labs-data-model.yml
idempotency:
  supported: false
  note: >-
    The Kits AI API does not document an idempotency-key mechanism. Create
    endpoints are not idempotent; each POST enqueues a new job.
versioning:
  scheme: uri-path
  current: v1
  base: https://arpeggi.io/api/kits/v1
  ref: lifecycle/arpeggi-labs-lifecycle.yml
rate_limiting:
  scope: account
  limits: 1000 GET/min, 1 POST/min
  throttled_status: 429
  ref: rate-limits/arpeggi-labs-rate-limits.yml
errors:
  envelope: plain-text body with conventional HTTP status codes
  ref: errors/arpeggi-labs-problem-types.yml