v2.0.0

What's Changed

• style(website): widen install code panels to match models table by @albrones in #2
• chore: remove dead code, dup files, and stale docs by @albrones in #3
• fix(engine): recover from wedged FSM so transcription never permanently hangs by @albrones in #4
• perf(bootstrap): cut ~75 MB runtime footprint and skip rebuild on re-run by @albrones in #5
• feat: adaptive transcription memory — learn from user corrections by @albrones in #6
• General audit: runtime bug fixes, hardening, CI gates, docs sync by @albrones in #7
• feat(asr)!: replace faster-whisper with NVIDIA Parakeet TDT 0.6b v3 by @albrones in #9
• fix(streaming): gate chunk boundaries on voiced speech, not elapsed time by @albrones in #12
• feat(trigger): drop the ⌃⌥⌘E/F presets and fix UTF-8 clipboard hand-off by @albrones in #11
• feat(trigger): add a toggle dictation mode by @albrones in #10
• feat(streaming): type each chunk as it lands in toggle mode by @albrones in #13
• chore(vercel): deploy only on merge to main, skip preview builds by @albrones in #14
• fix(audio): bound every model call to 30 s of audio by @albrones in #15
• chore: drop graphify, refresh the opsx tooling by @albrones in #17
• fix: dictation text fidelity, trigger strand, and config persistence by @albrones in #16
• fix(vercel): stop skipping production deploys by @albrones in #18
• docs(openspec): sync the four shipped changes into the main specs by @albrones in #19
• fix(engine): report live typing being disabled by streaming_enabled by @albrones in #20
• fix(linux): make the first run survivable, and clean up for release by @albrones in #21

New Contributors

• @albrones made their first contribution in #2

Full Changelog: v1.0.0...v2.0.0

v1.0.0

Full Changelog: v0.1.0...v1.0.0

v0.1.0

Full Changelog: https://github.com/albrones/whispy/commits/v0.1.0