The hold-to-talk release path called AsyncStream.Continuation.finish()
while holding MacAudioRecorder's NSLock; onTermination re-entered the
same lock on the main thread and froze the app. Finish snapshot sinks
outside the lock and guard the audio tap once stop() begins.
The dictation HUD no longer reassigns its hosting view and forces a
synchronous relayout on every view-model tick; it uses a fixed panel
size and SwiftUI @ObservedObject refresh instead.
Adds OSGKeyboardMacTests with a regression test that fails under the
old lock re-entry pattern.
Harden polish so question drafts stay questions, add local density
routing with style-specific degrade, expand fun style packs, and add
end-to-end FlowTrace logging plus offline guard eval scripts.
Unify Bailian/Volcengine/OpenAI realtime streaming, ABE polish routing with
fun styles, and a shared card-page Settings hierarchy; bump to 1.1 (build 32).
- Add FlowKeepAliveMode (Dynamic Island vs PiP) with iCloud sync
- Settings: keep-alive picker; inactivity timeout only for Live Activity
- PiP: waveform sample-buffer controller, mic released between utterances
- PiP sessions have no idle expiry; user closing PiP ends the session
- FlowSessionManager branches hostReady, session start, and utterance paths
- UIBackgroundModes picture-in-picture; bilingual strings and tests
Co-authored-by: Rocky <hkgood@users.noreply.github.com>
- Accumulate utterance PCM in FlowContinuousCapture for batch retry
- Run full-utterance transcribeChunk when stitched final lags partial
- Refactor Mac MLX tail drain to shared FlowUtteranceEndCoordinator
- Add FlowUtterancePCMStore, UtteranceBatchFallbackPolicy, and tests
Co-authored-by: Rocky <hkgood@users.noreply.github.com>
- Add FlowUtteranceEndCoordinator with 350ms silence drain and 150ms post-roll
- Extend FinalChunkRecovery for short/empty final chunks in chunked pipeline
- Snapshot partial at mic stop and guard final transcript in FlowSessionManager
- Unify tail drain presets (iosFlow/macMLX) and expand diagnostics
Co-authored-by: Rocky <hkgood@users.noreply.github.com>
Add shared prompt composition, custom style management, iCloud sync, and native iOS/macOS selection interfaces so users can keep a consistent writing voice across devices.
Use compliant permission prompts, align cloud ASR privacy disclosures, and default Flow sessions to a shorter inactivity window for safer resubmission.
Replace the Sherpa offline pipeline with native MLX streaming, resilient model downloads, live transcript previews, and supporting tests and documentation.
Complete phased plan for migrating macOS local dictation from Sherpa
offline CLI to mlx-audio-swift streaming with LocalASRBiasAdapter
vocabulary support.
Co-authored-by: Rocky <hkgood@users.noreply.github.com>
Replace the five-card stack with a pill tab bar and elevated detail
stage, use a white band for the competitor matrix, and a soft gray
band for the three-step setup section.
Replace the duplicate BYOK block with hover/tap detail panels under
the differentiators, drop the hero title period, center the matrix
footnote, and alternate section washes so vs / steps read as separate bands.
Retitle the differentiation, BYOK, and setup sections with richer
copy, and add a Typeless / Superwhisper / Openless / OSGKeyboard table
covering pricing, on-device ASR, BYOK, and platforms.
Add System/Light/Dark preference for iPhone and iPad, restore the iPad
sidebar brand mark, ship a DEBUG seed-demo URL for rich placeholder data,
and refresh GitHub Pages with the device composite, full-bleed pale-green
hero wash, and transparent Mac screenshot chrome.
Capture Mac and iPhone product shots for zh/en × light/dark, store under
assets/screenshots/{lang}/{theme}/, and swap images when the page
toggles language or appearance (including system theme).
Add optional StoreKit support tip (features stay free), redesign the
landing page with zh/en + light/dark and product shots, and ship
iPhone/iPad/Mac App Store screenshot sets.
Fix cold-start overlay recursion that overflowed the main-thread stack when
recording began while the ready overlay was still up; also remove temporary
on-screen Flow DEBUG panels after the orange-mic investigation, and land the
macOS overlay/catalog/layout polish plus related Flow recovery hardening.
End Live Activities and release the audio session synchronously on
applicationWillTerminate, and continue macOS local-model install,
onboarding, and settings polish on this branch.
Adds a bundled local ASR model catalog for the macOS app with one-click
Sherpa Qwen3 / SenseVoice downloads (pause/resume, inline actions) and a
shared model storage directory used by MLX Qwen3. Fixes the light-mode
sidebar material and makes the menu-bar icon follow the system appearance
with a refreshed status mark. Renames the built product to OSGKeyboard.app.
Bumps version to 0.5.2 (build 19).
Introduce a standalone macOS menu-bar app (OSGKeyboardMac) that reuses the
platform-agnostic OSGKeyboardShared core: record -> cloud/local ASR -> polish
-> insert. Local mode uses Qwen3-ASR via mlx-swift-asr (macOS 15+, Apple
Silicon); iOS targets stay zero-SPM.
Harden iCloud sync for multi-device correctness:
- Per-field settings merge (appSettings.v2) so concurrent edits no longer
clobber each other's unrelated fields.
- Per-device usage statistics (G-Counter) that sum instead of max().
- Tombstoned dictionary/history merge so deletes propagate and entries can't
resurrect.
- API keys replicate via iCloud Keychain, never iCloud KVS JSON; pulling a
legacy blob without key fields no longer wipes local Keychain entries.
- Add a low-risk "Sync Now" action in Settings.
Fix Flow keyboard mic state: stay orange until the host publishes a real ready
contract, share a single MicVoiceAvailability gate, and self-heal stale
cross-process heartbeat jitter instead of getting stuck.
Extract shared storage (SpeechHistoryStore/UsageStatisticsStore,
ConfigurationStore) into OSGKeyboardShared and add tests for the new
sync/merge logic.
Mirror the personal dictionary through iCloud Key-Value Store with
deterministic merge rules, main-app-only sync UI, and App Group as the
keyboard runtime cache. Add cloud-engine ASR with dictionary bias and
expand the bundled custom language model lexicon.
- fix(keyboard): drop unusable extensionContext.open; report the real
result of UIApplication.open and cancel the start watchdog immediately
on failure with clear "open OSGKeyboard" guidance instead of a 30s spin
- feat(capture): recover from audio route changes (AirPods/wired headset
plug-unplug) and phone-call interruptions by rebuilding the engine/tap
against the live hardware format
- feat(home): add a "Start" button in the Home footer when a session is
inactive and permissions are granted (no-jump manual restart)
- feat(hostapps): expand the return whitelist to 46 apps (Signal, Zoom,
Instagram, X, Douyin, Zhihu, Bilibili, Things, Firefox, ...) with
en/zh-Hans names and synced LSApplicationQueriesSchemes
- feat(island): use a "ready" checkmark instead of a mic glyph while idle
and set a staleDate so an orphaned Live Activity fades after a force-quit
- feat(coldstart): use the template OSG brand mark instead of a system
waveform glyph in the cold-start overlay
- perf(keyboard): rely on Darwin notifications as the primary session
signal and drop the fallback poll from 1s to 3s
- fix(keyboard): remove the redundant "Start" text button next to the
inactive-session hint on the keyboard
- docs: correct the per-take cap to 210s (3.5 min) and document that
force-quitting no longer resurrects the session (tracker + READMEs)
- Add OSGKeyboardLiveActivity widget extension (ActivityKit) with branded
compactLeading mark in the Dynamic Island during Flow sessions
- Wire FlowLiveActivityController into FlowSessionManager lifecycle
- Set keyboard PrimaryLanguage to mis in project.yml, Info.plist, and at runtime
- Enable NSSupportsLiveActivities on main app and widget extension targets
Co-authored-by: Rocky <hkgood@users.noreply.github.com>
- P0/P1: tail-drain state machine in FlowContinuousCapture with converter
flush; FlowSessionManager awaits drain before finalize
- P2: FlowCaptureTailDrain policy/tracker, pipeline diagnostics, unit tests
- P3: short final chunk merged re-transcription in ChunkedUtterancePipeline
- P4: UtteranceTranscriptStitcher composedSafely fallback; preview path parity
in LiveDictationController
Co-authored-by: Rocky <hkgood@users.noreply.github.com>
Reduce perceived latency from key release to final text:
- Adaptive chunking: 2.5s first chunk + 5s follow-ups so short
utterances start on-device recognition while still recording.
- Session-level ASR warmup and audio-format cache reuse to remove
per-utterance cold-start of SpeechAnalyzer.
- Mirror live pipelined partials to the keyboard transcript line via
a new flow.transcriptionPartial App Group key + Darwin ping.
Also commits the accumulated custom language model, Flow session,
keyboard extension restructure, and Xiaomi MiMo provider work in
progress on this branch.
Rework the marketing site with a cleaner, more restrained layout, replace emoji with Material Symbols monochrome icons, and add full SEO (Open Graph, Twitter cards, JSON-LD, canonical, robots.txt, sitemap.xml).
Rework the marketing site with a cleaner, more restrained layout, replace emoji with Material Symbols monochrome icons, and add full SEO (Open Graph, Twitter cards, JSON-LD, canonical, robots.txt, sitemap.xml).
Rocky