Skip to content

Changelog

Changelog

0.7.0 (2026-09-06)

Features

  • auto-download missing whisper.cpp models (#99) (f0d98ea), closes #71

Bug Fixes

  • stop auto-writing to ./output/, follow Unix stdout-default norm (#108) (9219f2a)
  • whisper.cpp engine parses -oj JSON instead of scraping console output (#97) (67e6acd)

0.6.1 (2026-08-30)

Bug Fixes

  • bound ffmpeg shutdown and stop live transcription racing temp-dir cleanup (#88) (0d06b07), closes #55
  • correct logging config precedence, record mutation, and unprefixed env vars (#59) (#94) (be6595b)
  • drain ffmpeg's stderr in live mode to prevent deadlock (#85) (d6190f4), closes #52
  • drop dead RuntimeError clause and avoid double pactl call in pulse.py (#96) (86156e3), closes #61
  • gate publishing on tests and stop dropping releases via path filters (#92) (9a65243), closes #57
  • only create output/ directory when it will actually be used (#95) (6ef7a1e), closes #60
  • raise on ffmpeg failure in record_once instead of discarding it (#87) (08e1cd0), closes #54
  • validate --output path is writable before recording or transcription (#56) (#90) (e3b46b2)
  • warn on --device/--engine combinations that silently fall back to CPU (#58) (#93) (0d37163)

0.6.0 (2026-08-30)

Features

  • add a public library API and make live capture a generator (#73) (038e4c7), closes #62
  • add SRT, WebVTT, JSON and TSV transcript output formats (#76) (89c9425)
  • report and cap how far live transcription falls behind recording (#79) (89b97e6), closes #66

Bug Fixes

  • tell transcription failures apart from silence in live mode (#75) (0860cbb), closes #53

Documentation

  • remove duplicate changelog entries (0f71e3f)
  • simplify security policy, remove outdated version table and email contact (154a236)

0.5.4 (2026-08-27)

Bug Fixes

  • handle Ctrl-C gracefully at top level and inform live-mode users (#45) (262d027)

0.5.3 (2026-08-27)

Bug Fixes

  • use PAT for release-please to trigger downstream publish workflow (#48) (01419bd)

0.5.2 (2026-08-27)

Bug Fixes

  • default log level to WARNING to suppress engine INFO spam (5268c21)

0.5.1 (2026-07-23)

Bug Fixes

  • always request timestamps from whisper-cli to fix empty transcripts (a3c36e7), closes #42

0.5.0 (2026-06-04)

Features

  • add --version flag to CLI (4ac9e20)

0.4.0 (2026-04-14)

Features

  • add ANSI color formatting to log output when stderr is a TTY (dc763e2)
  • add publish to pypi workflows (d3c0372)
  • add release-please for automatic version bumping (3e13457)
  • add security policy (6d48abd)
  • always write transcripts to output/ directory (4c76d52)
  • auto-stop live mode after a period of silence (152398d)
  • make whisper engines optional dependencies (db6f33c)
  • replace print statements with logging (dba30bc)
  • restrict workflow triggers with path filters (7756e4e)
  • thread-safe model caching in transcribe.py (221ccdb)

Bug Fixes

  • cache whisper models and add transcription timeout to prevent freezing (0f05e63)
  • don't specify model list (aa34c22)
  • prevent freezing on in-progress segments and add whisper.cpp engine (3df1991)
  • replace bare except clauses with specific exception types (f24cc63)
  • use explicit import check for faster-whisper auto-detection (1b22dab)
  • use pypa/gh-action-pypi-publish (9f86df4)

Documentation

  • add link to guide (eaad37c)
  • tweak README (eb8f1b1)
  • Update copyright year in LICENSE file (151b94f)
  • update installation instructions to reference PyPI and remove cd step (99314ee)
  • Update supported versions in SECURITY.md (8f150f8)

0.3.0 (2026-04-14)

Features

  • add ANSI color formatting to log output when stderr is a TTY (dc763e2)

0.2.0 (2026-04-09)

Features

  • add publish to pypi workflows (d3c0372)
  • add release-please for automatic version bumping (3e13457)
  • add security policy (6d48abd)
  • always write transcripts to output/ directory (4c76d52)
  • auto-stop live mode after a period of silence (152398d)
  • make whisper engines optional dependencies (db6f33c)
  • replace print statements with logging (dba30bc)
  • restrict workflow triggers with path filters (7756e4e)
  • thread-safe model caching in transcribe.py (221ccdb)

Bug Fixes

  • cache whisper models and add transcription timeout to prevent freezing (0f05e63)
  • don't specify model list (aa34c22)
  • prevent freezing on in-progress segments and add whisper.cpp engine (3df1991)
  • replace bare except clauses with specific exception types (f24cc63)
  • use explicit import check for faster-whisper auto-detection (1b22dab)
  • use pypa/gh-action-pypi-publish (9f86df4)

Documentation