generated: '2026-08-13'
method: searched
source: https://www.lalal.ai/changelog/
provider: LALAL.AI
provider_id: lalal-ai
url: https://www.lalal.ai/changelog/

scheme:
  style: date-based
  format: YYYY-MM-DD
  semver: false
  categories: [New Features, Improvements, Fixes]
  scope: >-
    Product-wide. Web app, pricing, payments, mobile apps, neural-network releases and the Developer API
    all share one stream; API changes are not separately tagged or filterable.
  breaking_change_labelling: false

current:
  api_version: v1
  api_spec_version: 1.1.0
  latest_entry: '2026-05-28'
  total_dated_entries: 131
  earliest_dated_entry: '2024-08-01'
  captured: '2026-08-13'
  note: >-
    131 dated entries were present on the page when captured. The window below is the API- and
    engine-relevant subset — the rest is billing, payments, localisation and UI work.

entries:
  - date: '2026-05-07'
    type: new
    highlights:
      - Added the new Lynx neural network to Voice Cleaner.
  - date: '2026-03-24'
    type: new
    highlights:
      - Added drum and bass stem support for the Andromeda splitter.
  - date: '2026-03-17'
    type: improvement
    highlights:
      - Removed legacy license-key sign-in for profile access; account login is now required.
      - Moved the Developer API documentation link to the site header so it is easier to find.
  - date: '2026-02-17'
    type: new
    breaking: false
    api: true
    highlights:
      - Added clear, structured error responses to the Developer API.
    note: >-
      This is the entry behind the {detail, code} envelope and the TaskStartErrorCodes /
      LicenseErrorCodes enums captured in errors/lalal-ai-problem-types.yml.
  - date: '2026-01-21'
    type: new
    api: true
    highlights:
      - Launched the new Developer API with interactive documentation.
  - date: '2026-01-16'
    type: new
    api: true
    highlights:
      - Published a machine-readable OpenAPI spec and an llms.txt file for the Developer API.
      - Updated the deprecation notice for the old API (v0) with links to the new Developer API documentation.
    note: >-
      LALAL.AI publishing its own llms.txt is a provider decision, not an inference — this entry dates it.
  - date: '2025-12-29'
    type: improvement
    api: true
    highlights:
      - Added a deprecation notice for the legacy API in the help response.
      - Fixed type conversion for task progress calculation.
  - date: '2025-12-22'
    type: new
    api: true
    breaking: false
    highlights:
      - Release of API v1.
      - 'Human-readable documentation: https://www.lalal.ai/api/v1/docs/'
      - 'OpenAPI specification: https://www.lalal.ai/api/v1/openapi.json'
  - date: '2025-11-21'
    type: new
    highlights:
      - Added Andromeda network support for Vocal and Instrumental and Voice and Noise stems.
  - date: '2025-11-17'
    type: new
    highlights:
      - Added the new Andromeda neural network for better vocal isolation, faster processing, cleaner stems.
  - date: '2025-09-30'
    type: new
    highlights:
      - Introduced a new Perseus-based separator for lead and backing vocals.
      - Added an enhanced dereverberation algorithm using the proprietary Lyra technology.
      - Added a new supported stem, Demuser — a Lyra-based model separating music from non-music components.
    note: The origin of the /api/v1/split/demuser/ and /api/v1/split/batch/demuser/ operations.
  - date: '2025-02-28'
    type: new
    api: true
    highlights:
      - Added Voice Cloner API to the Public API.
    note: The origin of /api/v1/change_voice/ and /api/v1/voice_packs/list/.
  - date: '2025-02-04'
    type: new
    highlights:
      - Added Electric Guitar and Piano stems via Perseus with Clear Cut and Deep Extraction settings.
      - Voice conversion can estimate and optionally apply echo and room tone from the original voice.
  - date: '2024-12-17'
    type: new
    highlights:
      - Added Acoustic Guitar stem separation on Perseus with Clear Cut and Deep Extraction modes.
  - date: '2024-11-22'
    type: new
    highlights:
      - Drums and Bass stems supported by the Perseus neural network.
      - Lead & Back Vocal Splitter now returns four tracks.
  - date: '2024-09-25'
    type: new
    highlights:
      - Released the Perseus transformer neural network for Vocal/Instrumental and Voice/Noise stems.

cross_references:
  - lifecycle/lalal-ai-lifecycle.yml