Skip to content

CHANGELOG.md

Latest commit

 

History

History
120 lines (80 loc) · 7.7 KB

CHANGELOG.md

File metadata and controls

120 lines (80 loc) · 7.7 KB

Changelog

2.5.1 (2026-04-26)

Bug Fixes

  • extend lmnt.lib path for sibling distributions (f848612)

2.5.0 (2026-04-26)

Features

  • allow lmnt.* subpackages from other distributions (8efe4fe)

2.4.1 (2026-04-26)

Chores

  • drop python 3.8 fallbacks and rye references (6138efa)
  • fix import ordering in generated types (094c4b9)
  • generate api_resources tests from carbonsteel (b9262db)
  • generate api.md from carbonsteel (8d72de7)
  • generate websocket types and sessions from asyncapi.yaml (5150f43)
  • migrate from rye to uv (a845166)
  • regenerate from new code generator (c4bb01b)
  • regenerate from new code generator (cc0b428)
  • register runtime files in manifest (e117e4b)
  • remove devcontainer config (c1da477)
  • remove MIGRATING.md (9bbb761)
  • remove SECURITY.md (22341da)
  • remove stainless sync state files (380b363)

v2.4.0 (2025-09-30)

Full Changelog: v2.3.0...v2.4.0

  • feat(api): api update (6e0a4dd)
  • feat(api): api update (cc7f909)
  • chore(internal): update pydantic dependency (a4c5af5)
  • chore(types): change optional parameter type from NotGiven to Omit (7c03fb8)
  • chore: do not install brew dependencies in ./scripts/bootstrap by default (75d2397)
  • chore: improve example values (5c0919c)

v2.3.0 (2025-09-03)

Full Changelog: v2.2.1...v2.3.0

  • replace ava with leah (47bb854)
  • avoid newer type syntax (773b08e)
  • chore(internal): update pyright exclude list (8c1bd1d)
  • feat(api): api update (7820c00)
  • chore(internal): add Sequence related utils (f1f239b)

v2.2.1 (2025-08-21)

Full Changelog: v2.2.0...v2.2.1

  • chore(internal): fix ruff target version (08c3067)
  • chore: update @stainless-api/prism-cli to v5.15.0 (1634751)
  • chore(internal): update comment in script (6649334)

v2.2.0 (2025-08-05)

Full Changelog: v2.1.0...v2.2.0

  • feat: add support for resets in websocket speech sessions

v2.1.0 (2025-08-05)

Full Changelog: v2.0.0...v2.1.0

  • fix(parsing): ignore empty metadata (2671330)
  • fix(parsing): parse extra field types (e652a62)
  • chore(project): add settings file for vscode (9d450b1)
  • feat(api): api update (f0ef880)

July 17, 2025

2.0.0

  • BREAKING CHANGES: The new v2 SDK provides more streaming functionality, a more modern, type-safe interface with better error handling, and improved performance. To migrate from the legacy v1 SDK, please update your code to use the new behavior or pin to a previous version if preferred. More details in the migration guide.

1.1.0

Jan 3, 2024

  • synthesize_streaming will now return a buffer_empty boolean when extras are requested. This can be used to determine when the server has no more audio to send after the client has sent a flush message.
  • A bug that caused create_voice to fail in Python 3.8 and earlier has been fixed.

1.0.1

Nov 18, 2023

  • Changes the owner type in list_voices from 'lmnt' to 'system' to match the LMNT REST API.

1.0.0

Nov 14, 2023

  • Uses the stable v1 LMNT REST API.
  • BREAKING CHANGES: please update your code to use the new behavior or pin to a previous version if preferred.
    • Default audio encoding format in synthesize is now mp3 (previously wav). Format can be specified by adding format='wav' or format='mp3' to the synthesize call.
    • list_voices now returns a list of voice dictionaries for simplicity of return values and ease of use. Previously it returned a dictionary with key voices which contained a dictionary of voice dictionaries keyed by their voice id.
    • synthesize no longer returns just the binary audio data. It instead always returns a dictionary with keys audio, durations (optional), and seed (optional).
    • synthesize_streaming no longer returns an aiohttp WSMessage. It instead returns a dictionary with keys audio, durations (optional), and seed (optional).
  • Adds support for creating, updating, and deleting voices as well as getting account details.

0.2.1

Oct 12, 2023

  • Removes the default seed of synthesize (previously 0). When seed is unspecified, synthesize will now use a random seed.
  • The durations dictionary has been updated. Now includes each word itself and changes the units of duration from the number of samples to seconds.

0.2.0

Sep 7, 2023

  • Add support for durations.

0.1.0

Aug 11, 2023

  • Initial release