feat(audio): desktop Silero ONNX VAD + Windows PTT modernization + MSVC CRT build fix (#37)

* feat(audio): add Silero ONNX VAD with WebRTC fallback

Introduce SileroOnnxVad and SileroOnnxVadWorker for desktop targets. The worker runs Silero v6 ONNX inference on a dedicated thread, accumulating 10 ms frames into the 512-sample 16 kHz input the model expects. Add VadOutput, VoiceActivityDetector trait, and WebRtcFallbackVad to provide a uniform VAD interface with graceful fallback when the ONNX model is unavailable. Wire the new VadBackend variants through AudioProcessingConfig and the snapshot stats so the bridge can report which detector is active.

* feat(audio): integrate desktop VAD worker into capture engine

Wire SileroOnnxVadWorker into the desktop capture path so voice activity can open the transmit gate before encoding. The capture callback now processes all audio through resample, downmix, and VAD unconditionally; transmit_active still gates Opus encoding.

Add new_desktop_audio_processing_state() to construct the config/stats/worker triple, and apply_desktop_vad_backend() to synchronously load or clear the worker on config changes. Override processing_backend to Noop for desktop so bridge diagnostics report the correct backend rather than the iOS-oriented PlatformVoiceProcessing default.

Includes review-driven cleanups: StreamConfig clone to deref per clippy, and a comment explaining why two try_lock calls on silero_vad_worker are structurally necessary (borrow checker requires the policy probe and the fallback path to not share a lock guard because mark_vad_fallback_active takes &mut self).

* fix(audio): modernize Windows PTT to current windows-rs API

Port the Raw Input plus low-level keyboard hook PTT backend to the newer windows-rs patterns: OptionalHandle, Result-returning CreateWindowExW, and None for CallNextHookEx. Replaces the old HHOOK(0) pointer casts. Add deterministic tests for mouse button 4 and 5 press and release driving the gate.

* build(windows): force MSVC release CRT for audiopus cmake builds

audiopus_sys calls cmake::build(opus_path), so downstream Cargo env cannot use cmake-rs Config::define() to override CMake's MSVC Debug CRT defaults. Point cmake-rs at a small wrapper that injects the policy and cache variables during configure while passing cmake --build, --version, and -E through unchanged. Keeps Opus Debug builds on Rust's release dynamic CRT (/MD) instead of CMake's default debug CRT (/MDd), which otherwise pulls in unresolved __imp__CrtDbgReportW symbols at test link.

Document that the iOS deployment target is intentionally absent from this file. It is enforced by tools/build-ios.sh and the Xcode project; setting it globally here would make native macOS cargo check runs try to link iPhone objects against the macOS SDK.

* build(flutter): update pubspec.lock after plugin additions

Regenerated lockfile reflecting the local_notifications and connectivity_plus plugin additions from the poke-notifications feature.

* fix(audio): address PR #37 review findings

Six fixes from independent PR review:

1. BLOCKER: Replace Windows-only cmake .cmd wrapper with cross-platform
   CMake env vars. Setting CMAKE=tools/cmake-msvc-release-crt.cmd
   globally broke non-Windows hosts because cmake-rs would try to
   execute a .cmd file on macOS/Linux. Instead, set
   CMAKE_POLICY_DEFAULT_CMP0091=NEW and CMAKE_MSVC_RUNTIME_LIBRARY=
   MultiThreadedDLL as env vars that CMake reads natively. MSVC-
   specific vars are safely ignored by GCC/Clang toolchains. Delete
   the now-unnecessary wrapper script.

2. IMPORTANT: Join the Silero worker thread in Drop instead of
   detaching it. The old code dropped the JoinHandle which detaches
   the thread; the new code calls handle.join() after closing the
   channel, ensuring the ONNX session is cleaned up before the
   worker is replaced during config changes.

3. IMPORTANT: Single-try_lock refactor of the capture VAD callback.
   The double try_lock (policy probe + send) is replaced by a single
   scoped try_lock that both probes availability and sends the frame.
   The guard is dropped before the fallback path, which needs &mut
   self for mark_vad_fallback_active. This also eliminates the
   VadWorkerPolicy enum and callback_vad_worker_policy function,
   whose behavior is now inlined into the callback.

4. IMPORTANT: Remove tracing from the realtime capture callback.
   mark_vad_fallback_active and sync_vad_backend emitted info!/warn!
   from the audio thread. Replace with silent atomic state
   publishing via SharedAudioProcessingStats; the bridge stats
   stream already exposes vad_fallback_active for diagnostics.

5. IMPORTANT: Defer ONNX model load outside the worker mutex.
   apply_desktop_vad_backend_to_worker now constructs the new worker
   before taking the lock, then swaps it in under a short hold.
   This prevents the realtime callback from being blocked during
   model I/O + thread spawn.

6. MINOR: Remove unused VadBackend import from vad/mod.rs after
   deleting the policy code.

* fix(audio): address PR #37 second-pass review findings

5-agent review found 5 blocking issues. All addressed:

1. BLOCKER: CMake env vars don't reach CMake cache. Restored .cmd wrapper
   but scoped to Windows MSVC targets only via [target.x86_64-pc-windows-msvc]
   and [target.aarch64-pc-windows-msvc] in .cargo/config.toml. Non-Windows
   hosts are unaffected.

2. BLOCKER: processing_backend normalized in set_audio_processing_config
   on desktop (cfg-gated override to Noop), mirroring startup default.

3. BLOCKER: Model-path reload was already wired via reload_audio_processing_config.
   Fixed misleading doc comment in core/lib.rs.

4. BLOCKER: DEC-030 updated to reflect desktop VoiceActivity enablement.
   Traceability docs (SRS, SysDes, SAD, SDD, implementation-status) updated.

5. Silero ONNX cfg narrowed to desktop-only (excludes macOS/Android).
   Cargo.toml ort dependency target cfg narrowed similarly.

6. Realtime callback debt documented as TODO at CaptureState::ingest.

* fix(audio): exclude ort dep on Android target

ort does not provide first-class Android prebuilts in our pin, mirror the
iOS/macOS exclusion so cargo metadata succeeds for android targets.

* test(audio): fix stale select_ptt_backend import in ptt_privacy

The helper moved out of the ptt_backends submodule onto the crate root;
update the integration test imports so the test compiles again.

* build(windows): scope MSVC release CRT cmake wrapper via Cargo [env]

Cargo's [target.<triple>] table only forwards a fixed allowlist
(linker, runner, rustflags, rustdocflags, ar), so setting CMAKE there
was silently dropped and audiopus_sys kept linking the debug CRT,
producing LNK4098 'MSVCRTD conflicts' and __imp__CrtDbgReportW errors
on x86_64-pc-windows-msvc test builds.

Move the override to Cargo's [env] table using cc/cmake-rs's
target-suffixed CMAKE_<triple> lookup (force=true, relative=true) so it
applies to MSVC targets only and not to host tooling. Add stdout
markers to the wrapper so its invocation is provable in cargo -vv logs.

Verified: cargo test -p chanora_audio --target x86_64-pc-windows-msvc
--lib --no-run now links cleanly; CMakeCache.txt records
CMAKE_MSVC_RUNTIME_LIBRARY=MultiThreadedDLL and CMP0091=NEW.

* fix(flutter): gate VoiceActivity transmit mode by platform support

VoiceActivity relies on the native VAD worker, which is only wired up
on Windows, Linux, and Android. Showing the option on iOS, macOS, or
web let users select a mode that silently never transmitted.

Add voiceActivityTransmitAvailable + transmitModeSegmentsFor() helpers
in voice_settings_controls.dart, hide the VAD row in voice_compact.dart
and drop the VAD segment from the settings dialog when unsupported.
Keep the legacy const transmitModeSegments for the existing widget test
and add two new tests covering the gated helper.

docs/architecture/sad.md

@@ -123,7 +123,7 @@ Runtime event or error
 | Bridge command DTOs | Flutter generated API | `chanora_bridge`, Rust core | Stable typed DTOs; no raw protocol-library types cross to Flutter |
 | Bridge event DTOs | Rust core / bridge | Flutter services/widgets | User-safe errors and capability fields are explicit |
 | Protocol DTOs | `chanora_protocol` | Rust core, state sync | Protocol adapter isolates `tsclientlib` |
-| Audio configuration | Flutter settings / Rust core | `chanora_audio` | Voice modes and processing flags are explicit; VAD remains disabled/deferred |
+| Audio configuration | Flutter settings / Rust core | `chanora_audio` | Voice modes and processing flags are explicit; Windows/Linux desktop VAD-backed `VoiceActivity` is enabled only where runtime evidence exists, with unsupported platforms disabled/deferred |
 | Storage records | Storage crate | Rust core / Flutter UI via bridge | Secrets stay behind secure-storage abstraction |
 | Diagnostic bundles | Diagnostics crate | Flutter diagnostics UI | Redaction runs before export or display |
 | Platform capability records | Platform adapters/audio/PTT backends | UI and release record | UI/release wording must not over-claim capability |
@@ -159,7 +159,7 @@ Runtime event or error
 | Secure storage abstraction | Platform storage details do not leak into UI or unrelated crates |
 | Advisory audio benchmarks | Performance regressions are surfaced without making CI a hard release gate at this stage |
 | PTT capability levels | Platform PTT support is represented as capability data and must match release wording |
-| VoiceActivity deferral | `VoiceActivity` remains reserved/disabled until a later baseline allocates implementation |
+| VoiceActivity platform scope | Windows/Linux desktop `VoiceActivity` is implemented through the capture VAD path; unsupported platforms remain disabled/deferred until backend allocation and runtime verification exist |
 | No automatic diagnostic upload in MVP | Diagnostics are local and user-initiated unless future approved requirements change policy |
 
 ## 11. Verification Handoff

docs/architecture/sdd.md

@@ -63,7 +63,7 @@ Design rules:
 | Capture/playback | Platform-specific units handle Android, iOS, desktop/fallback paths behind Rust audio abstractions |
 | Codec | Opus encode/decode lives in `opus_voice.rs` and associated audio modules |
 | DSP chain | High-pass filter, noise suppression, echo cancellation, and AGC are represented by audio processing modules/backends |
-| Transmit control | `TransmitMode` supports `Ptt`, `Continuous`, and reserved `VoiceActivity`; `VoiceActivity` has no active MVP implementation |
+| Transmit control | `TransmitMode` supports `Ptt`, `Continuous`, and `VoiceActivity`; `VoiceActivity` is active for Windows/Linux desktop capture when VAD is configured, while mobile, macOS, and unverified-platform enablement remain deferred |
 | VoiceActivity gate (capture-side) | `voice_activity::VoiceActivityStateMachine` is the 10 ms-cadence gate for `TransmitMode::VoiceActivity`; open-after 40 ms (debounce), hangover 500 ms (anti-chatter), min-tx 200 ms (anti-flicker), weak-hold 30-100 frames (anti-stale-VAD); live `configure()` re-clamps existing timers on settings change without resetting state; 9 unit tests cover the main paths |
 | PTT | Desktop/mobile backends expose capability level and active backend; missed-key-up watchdog prevents stuck transmit |
 | Release tail | Tail handling prevents abrupt cutoffs after PTT release where configured |

docs/governance/decision-impact-assessment.md

@@ -11,7 +11,7 @@
 | Android Keystore-backed DEK deferred | Secure-storage claim limited | Storage design carries fallback limitation | Platform audit required | Waiver required |
 | Full reducer tests incomplete | State verification incomplete | State design remains valid but evidence partial | SWE.4/SWE.5 partial | Blocks full state-sync claim |
 | Desktop/iOS artifacts not release-ready | Platform packaging requirements partial | Release design remains source-build/unsigned | SYS.4 evidence partial | Public binary release No-Go |
-| VAD deferred | VoiceActivity not active | UI must show disabled/coming-soon | No VAD pass claim | No VAD marketing claim |
+| VAD platform-scoped | VoiceActivity active only for verified Windows/Linux desktop paths | UI must show disabled/unavailable on unsupported platforms | VAD pass claim must name verified platform/runtime evidence | No broad VAD marketing claim without platform scope |
 
 ## 2. Conclusion
 

docs/governance/maintainability-review-2026-06-08.md

@@ -43,7 +43,7 @@ This record captures the current maintainability review so implementation, verif
 | Android Keystore-backed DEK remains deferred | Android identity/bookmark encryption has weaker fail-safe properties than final target secure-storage design. | Android secure-storage audit or waiver; explicit release-readiness limitation. |
 | Android permission/audio lifecycle needs deeper route exercise | Build/install/launch smoke now passes on the emulator, but full permission-flow and audio-route lifecycle behavior still need an interactive scenario or device test before release. | Device/emulator scenario covering permission request/denial/grant, voice controls, foreground service, audio focus, and route/SCO transitions. |
 | iOS device runtime verification not executed in this review | The iOS audio-session error path is hardened, but VoiceProcessingIO/session ordering and runtime audio behavior still need device evidence. | iOS device or simulator build/run plus audio-session smoke evidence before iOS runtime success is claimed. |
-| VAD / VoiceActivity wording drift | VAD assets, tests, and scaffolding exist, but product-enabled `VoiceActivity` remains reserved/disabled per DEC-030. | Release, README, and verification wording must distinguish scaffolding/assets/tests from shipped product behavior. |
+| VAD / VoiceActivity wording drift | VAD assets, tests, and Windows/Linux desktop runtime wiring exist, but product-enabled `VoiceActivity` must remain platform-scoped per DEC-030. | Release, README, and verification wording must distinguish verified desktop behavior from unsupported mobile/macOS/unverified-platform behavior. |
 | Protocol voice packet re-export is an intentional exception | Future maintainers may assume complete protocol isolation and accidentally widen the Seam. | Architecture note in SAD/SDD or a decision-register entry. |
 | Bridge DTO mirror drift | Field additions can be missed across Core, Bridge, and Dart generated DTOs. | Bridge generation check plus Flutter analyze/test after bridge DTO changes. |
 | Full live reducer integration remains separate from reducer unit coverage | State reducer tests are strong, but runtime UI still has snapshot/probe paths. | SWE.5 integration run proving live protocol events fold through the intended state path, or explicit P1 deferral. |
@@ -91,7 +91,7 @@ adb -s emulator-5554 shell pidof app.chanora.chanora_flutter
 
 - Android minimum runtime baseline is API 28 (Android 9.0) per SysRS-288, SRS-187, DEC-004, and the Gradle `minSdk = 28` configuration. Documents must not revive the older API 24 baseline.
 - Flutter app version/build is `0.3.0+100` in `apps/chanora_flutter/pubspec.yaml`. Rust workspace package version remains `0.2.0-beta.1`. Release documents must distinguish these values instead of treating them as one candidate version.
-- The v0.3.0 changelog entry may mention VAD assets/backends only as implementation scaffolding; product-enabled `VoiceActivity` remains disabled/coming-soon until DEC-030 is superseded and runtime verification exists.
+- The v0.3.0 changelog entry may mention Windows/Linux desktop VAD-backed `VoiceActivity` only with matching runtime evidence; mobile, macOS, and unverified-platform `VoiceActivity` remain disabled/unavailable until DEC-030 is superseded and runtime verification exists.
 - README wording must describe the existing Flutter/Rust workspace and app scaffold, not a future scaffold that has not been created.
 
 ## 8. Git Policy

docs/governance/product-decision-register.md

@@ -15,7 +15,7 @@ This register records product and engineering decisions referenced by the DV doc
 | DEC-012 legal/trademark/OSS review | Open | Blocks public/store release |
 | DEC-020 dual license MIT OR Apache-2.0 | Accepted per README | Supports license posture; dependency notices still require review |
 | DEC-027 desktop mouse side-button PTT | Accepted by requirements baseline | Verification must not over-claim unsupported platform input classes |
-| DEC-030 VAD deferral | Accepted as deferral | VAD scaffolding, assets, and tests may exist, but product `VoiceActivity` remains disabled/coming-soon until a later baseline enables and verifies it |
+| DEC-030 VAD platform scope | Partially superseded by desktop enablement | Windows/Linux desktop `VoiceActivity` is enabled through the audio capture VAD path and must be claimed only with matching runtime evidence; mobile, macOS, and unverified-platform `VoiceActivity` remain disabled/deferred until a later baseline enables and verifies them |
 | DEC-032 Android CMake patch exit path | Active tracking | Patched dependency requires reevaluation |
 | DEC-033 macOS VPIO ducking configuration | Accepted | Write `kAUVoiceIOProperty_OtherAudioDuckingConfiguration` with `mEnableAdvancedDucking=0` (disables dynamic voice-activity-driven ducking) and `mDuckingLevel=Min` (= 10) to minimise the ducking of other apps' audio during a voice session; property is macOS 14+ only, the macOS 13 set fails silently (debug log) and VPIO uses its default behaviour; matches the iOS `.voiceChat` baseline on macOS 14+ |
 | DEC-034 Android runtime verification gate | Active tracking | Android target compilation, install, and runtime smoke are blocked locally until `aarch64-linux-android-clang` is available and `adb devices -l` shows an authorized target; release docs must not claim Android runtime success |

docs/implementation-status-2026-05-28.md

@@ -21,7 +21,7 @@
 | Protocol adapter | `chanora_protocol` — `tsclientlib` isolated behind `ProtocolClient`, typed DTOs, `ProtocolError` catalogue |
 | Connection lifecycle | `chanora_core` — supervisor task, exponential backoff reconnect (1s→60s), user-disconnect suppresses reconnect; branch `simplify-project-review` has started splitting the previous large `lib.rs` into focused internal Modules (`events.rs`, `network_diagnostics.rs`) while preserving public re-exports |
 | State sync reducer unit | `chanora_state` — `ConnectionState`, `channel_join`, snapshot/delta reducers, reconnect handling, deterministic ordering, malformed duplicate normalization, channel-delete/client cleanup, and reducer unit tests. Runtime core integration still uses snapshot/probe refresh paths and remains separate validation work. |
-| Audio subsystem | `chanora_audio` — Opus encode/decode, HPF/NS/AEC3/AGC2 DSP, PTT backends (Windows/macOS/Linux/focused), iOS VoiceProcessingIO, Android Oboe, jitter buffer via `tsclientlib::audio::AudioHandler`, mixer, mute/deaf gates, release-tail timer, and VAD scaffolding/assets. Product `VoiceActivity` remains disabled per DEC-030. |
+| Audio subsystem | `chanora_audio` — Opus encode/decode, HPF/NS/AEC3/AGC2 DSP, PTT backends (Windows/macOS/Linux/focused), iOS VoiceProcessingIO, Android Oboe, jitter buffer via `tsclientlib::audio::AudioHandler`, mixer, mute/deaf gates, release-tail timer, and Windows/Linux desktop `VoiceActivity` through the capture VAD path. Mobile, macOS, and unverified-platform `VoiceActivity` remain deferred per DEC-030. |
 | Push-to-talk | Per-platform backends: Windows Raw Input + hook fallback, macOS Event Tap, Linux freedesktop portal, focused fallback; `PttCapabilityLevel` (L0–L3); missed-key-up watchdog |
 | Voice controls UI | `voice_bar`, `voice_compact`, `voice_haptics`, `voice_level_meter`, `voice_platform`, `ptt_capability_badge`, `talk_power_warning` |
 | Storage (non-secret) | `chanora_storage` — `BookmarkRepository` (SQLite/rusqlite bundled, schema v2), ChaCha20-Poly1305 encrypted passwords |
@@ -54,7 +54,7 @@
 |---|---|
 | Event replay tooling | Reducer tests cover the state-sync contract, but standalone replay-file tooling remains a P1 verification gap. |
 | Reducer runtime integration evidence | The standalone reducer is unit-tested, but `chanora_core` still refreshes UI state through snapshot/probe paths rather than folding all live protocol events through `chanora_state::reduce`. |
-| Silero VAD | `assets/models/silero_vad.onnx` bundled but DEC-030 defers VAD to P1; `TransmitMode::VoiceActivity` is reserved and disabled in this baseline. |
+| Mobile/macOS VoiceActivity | `assets/models/silero_vad.onnx` is bundled and used by the desktop VAD path where runtime evidence supports it; mobile, macOS, and unverified-platform `TransmitMode::VoiceActivity` remain disabled/deferred until a later baseline supplies backend enablement and verification evidence. |
 | macOS build | Source-buildable only; no public release artifact is approved. |
 | Windows build | Source-buildable only; no public release artifact is approved. |
 | iOS build | Source-buildable/unsigned validation only; no TestFlight/App Store release artifact is approved. |

docs/release/dv-waiver-register.md

@@ -18,7 +18,7 @@ A waiver records a known gap that reviewers may accept for a limited decision sc
 | DV-WVR-004 | Standalone event replay and runtime reducer-integration evidence not yet complete | `docs/implementation-status-2026-05-28.md`, local `cargo test -p chanora_state --locked` evidence | Reducer unit coverage supports state synchronization, but file-based replay evidence and live-event integration evidence remain open | DV documentation pass and internal validation | Event replay tooling implemented or requirement reprioritized; runtime reducer integration evidence attached |
 | DV-WVR-005 | Event replay tool not found | `docs/implementation-status-2026-05-28.md` | Limits P1 state verification hooks SRS-061/SRS-098 | Accepted as P1 deferral | Event replay tooling implemented or requirement reprioritized |
 | DV-WVR-006 | Desktop and iOS artifacts are source-buildable or unsigned only | `docs/implementation-status-2026-05-28.md`, `docs/release/ios-build.md` | Blocks packaged public release claims | Internal validation from source/unsigned builds only | Signed/notarized/package artifacts exist and hashes are recorded |
-| DV-WVR-007 | Silero VAD asset bundled while `VoiceActivity` is deferred | `docs/implementation-status-2026-05-28.md` | Risk that UI/release wording overstates VAD availability | DV may pass if VoiceActivity remains disabled/coming-soon | VAD implementation allocated in a later baseline or asset/wording reconciled |
+| DV-WVR-007 | Silero VAD asset bundled while `VoiceActivity` is platform-scoped | `docs/implementation-status-2026-05-28.md` | Risk that UI/release wording overstates VAD availability beyond verified Windows/Linux desktop paths | DV may pass if VoiceActivity claims are limited to verified desktop evidence and unsupported platforms remain disabled/unavailable | Mobile/macOS implementation allocated in a later baseline or asset/wording reconciled |
 | DV-WVR-008 | Artifact hashes, tag, and candidate run IDs are not recorded in release record | `docs/release/release-readiness-go-nogo-record.md` | Blocks final release approval and reproducibility | DV documentation review only | Candidate build run records, tag, commit SHA, and artifact hashes are recorded |
 | DV-WVR-009 | Android target compile and runtime smoke blocked locally | `docs/governance/maintainability-review-2026-06-08.md`, `docs/release/release-readiness-go-nogo-record.md` | Blocks Android runtime, permission-flow, and audio-lifecycle success claims | Documentation review only; internal validation must keep Android limitation stated | Android NDK compiler `aarch64-linux-android-clang` is available, `adb devices -l` shows an authorized target, and Android build/install/smoke evidence is attached |
 

docs/srs.md

@@ -2597,7 +2597,7 @@ This section extends the ASPICE SWE.1 Software Requirements Specification. The s
 - Source SysDes: SysDes-150
 - Verification method: Integration Test, UI Review
 
-**SRS-205**: The software shall represent the user's voice transmit mode as a `TransmitMode` enum with variants `Ptt`, `Continuous`, and `VoiceActivity` (the last reserved with no v1 implementation per DEC-030). The setting shall be persisted per identity via the identity store. The default value for a fresh install shall be `Ptt`. The UI shall render `VoiceActivity` as a disabled "coming soon" option until an implementation is allocated in a later baseline.
+**SRS-205**: The software shall represent the user's voice transmit mode as a `TransmitMode` enum with variants `Ptt`, `Continuous`, and `VoiceActivity`. `VoiceActivity` shall be selectable only on platforms with an implemented and verified VAD capture path in this baseline (currently Windows/Linux desktop); mobile, macOS, and unverified-platform enablement remain deferred per DEC-030. The setting shall be persisted per identity via the identity store where the selected platform supports it. The default value for a fresh install shall be `Ptt`. The UI shall render `VoiceActivity` as disabled/unavailable on unsupported platforms rather than claiming runtime support.
 
 - Status: Baseline Candidate
 - Type: Software Interface Requirement
@@ -2837,7 +2837,7 @@ Consistent with the SysRS-307 / SysRS-308 / SysRS-309 deferrals propagated throu
 
 | Version | Date | Description |
 |---|---|---|
-| 0.9.5 | 2026-05-15 | Added v1 audio + PTT lifecycle software requirements SRS-204 through SRS-207 sourced from SysDes-149..151: bridge surface drops `start_audio` / `stop_audio` and adds `voice_join(channel_id)` / `voice_leave()`, audio engine opens streams on first voice-channel join and closes on last leave with output independent of mic-permission state, `TransmitMode` enum (`Ptt` default, `Continuous`, reserved `VoiceActivity` per DEC-030) persisted per identity, `release_tail_ms` (default 200, range 0–500) gate on the `true → false` transition of `transmit_active` with re-press cancellation, Voice Bar hard-mute override of `transmit_active`. Strict layered sourcing preserved (`SRS -> SysDes` only). |
+| 0.9.5 | 2026-05-15 | Added v1 audio + PTT lifecycle software requirements SRS-204 through SRS-207 sourced from SysDes-149..151: bridge surface drops `start_audio` / `stop_audio` and adds `voice_join(channel_id)` / `voice_leave()`, audio engine opens streams on first voice-channel join and closes on last leave with output independent of mic-permission state, `TransmitMode` enum (`Ptt` default, `Continuous`, platform-scoped `VoiceActivity` per DEC-030) persisted per identity where supported, `release_tail_ms` (default 200, range 0–500) gate on the `true → false` transition of `transmit_active` with re-press cancellation, Voice Bar hard-mute override of `transmit_active`. Strict layered sourcing preserved (`SRS -> SysDes` only). |
 
 
 ## Baseline Candidate 0.9.6 Update

docs/sysdes.md

@@ -2852,7 +2852,7 @@ This SysDes version covers all known SysRS requirements from `SysRS-001` through
 - ASPICE SYS.3 alignment: Architecture constraints
 - Allocated SysRS: SysRS-298
 
-**SysDes-149**: The system architecture shall define a `TransmitMode` element carried as an enum at the audio + bridge + UI boundary with variants `Ptt`, `Continuous`, and a reserved `VoiceActivity` placeholder that has no allocated implementation in this baseline (deferred per DEC-030). The active mode shall be persisted in the identity store; the UI shall surface `VoiceActivity` as a disabled "coming soon" option until an implementation is allocated in a later baseline.
+**SysDes-149**: The system architecture shall define a `TransmitMode` element carried as an enum at the audio + bridge + UI boundary with variants `Ptt`, `Continuous`, and `VoiceActivity`. `VoiceActivity` shall be surfaced only when the current platform has an allocated, implemented, and verified VAD capture path in this baseline (currently Windows/Linux desktop); mobile, macOS, and unverified-platform enablement remain deferred per DEC-030. The active mode shall be persisted in the identity store where the selected platform supports it; unsupported platforms shall surface `VoiceActivity` as disabled/unavailable rather than claiming runtime support.
 
 - Status: Baseline Candidate
 - Type: Subsystem Interface
@@ -3029,7 +3029,7 @@ This SysDes version covers all known SysRS requirements from `SysRS-001` through
 
 | Version | Date | Description |
 |---|---|---|
-| 0.9.5 | 2026-05-15 | Added v1 audio + PTT lifecycle allocation SysDes-149 through SysDes-151 sourced from SysRS-303 / SysRS-304: `TransmitMode` enum element (`Ptt` / `Continuous` / reserved `VoiceActivity` per DEC-030) at the audio + bridge + UI boundary, audio engine lifecycle bound to voice-channel membership with no manual start affordance and a listen-only path independent of mic permission, hard-mute override element, and the release-tail timer adapter (default 200 ms, range 0–500 ms) on `transmit_active`. Strict layered sourcing preserved (`SysDes -> SysRS` only). |
+| 0.9.5 | 2026-05-15 | Added v1 audio + PTT lifecycle allocation SysDes-149 through SysDes-151 sourced from SysRS-303 / SysRS-304: `TransmitMode` enum element (`Ptt` / `Continuous` / platform-scoped `VoiceActivity` per DEC-030) at the audio + bridge + UI boundary, audio engine lifecycle bound to voice-channel membership with no manual start affordance and a listen-only path independent of mic permission, hard-mute override element, and the release-tail timer adapter (default 200 ms, range 0–500 ms) on `transmit_active`. Strict layered sourcing preserved (`SysDes -> SysRS` only). |
 
 
 ## Baseline Candidate 0.9.6 Update

docs/sysrs.md

@@ -1924,7 +1924,7 @@ This section converts the baseline product decisions into auditable system-level
 - Priority: P0
 - Verification: Privacy Review, Security Audit, Diagnostic Inspection
 
-**SysRS-303**: The Chanora application system shall define the v1 voice transmit mode set as `Ptt` and `Continuous`, with `VoiceActivity` reserved on the enum surface but unimplemented in this baseline (deferred per DEC-030). The default mode on a fresh install shall be `Ptt`; the user-selected mode shall be persisted per identity. The audio engine lifecycle shall be bound to voice-channel membership: input and output streams shall open on the user's first voice-channel join of the session and shall close on the last voice-channel leave, with no manual start affordance exposed at any system interface. The output stream shall open regardless of microphone-permission state so listen-only is a first-class flow. A user-facing hard-mute toggle shall force the transmit gate closed and shall override the active transmit mode, the PTT key state, and every other internal signal.
+**SysRS-303**: The Chanora application system shall define the v1 voice transmit mode set as `Ptt`, `Continuous`, and `VoiceActivity`. `VoiceActivity` shall be enabled only on platforms with an implemented and verified VAD capture path in this baseline (currently Windows/Linux desktop); mobile, macOS, and unverified-platform enablement remain deferred per DEC-030. The default mode on a fresh install shall be `Ptt`; the user-selected mode shall be persisted per identity where the selected platform supports it. The audio engine lifecycle shall be bound to voice-channel membership: input and output streams shall open on the user's first voice-channel join of the session and shall close on the last voice-channel leave, with no manual start affordance exposed at any system interface. The output stream shall open regardless of microphone-permission state so listen-only is a first-class flow. A user-facing hard-mute toggle shall force the transmit gate closed and shall override the active transmit mode, the PTT key state, and every other internal signal.
 
 - Priority: P0
 - Verification: Functional Test, UX Review
@@ -1989,7 +1989,7 @@ This section converts the baseline product decisions into auditable system-level
 
 | Version | Date | Description |
 |---|---|---|
-| 0.9.5 | 2026-05-15 | Added v1 audio + PTT lifecycle requirements SysRS-303 and SysRS-304 capturing the no-manual-start audio engine bound to voice-channel membership, the v1 transmit-mode set (`Ptt` default + `Continuous`, with `VoiceActivity` reserved per DEC-030), the listen-only flow (output independent of mic permission), the hard-mute override, and the 200 ms (0–500 ms) PTT release tail for word-boundary anti-clipping. |
+| 0.9.5 | 2026-05-15 | Added v1 audio + PTT lifecycle requirements SysRS-303 and SysRS-304 capturing the no-manual-start audio engine bound to voice-channel membership, the v1 transmit-mode set (`Ptt` default, `Continuous`, and platform-scoped `VoiceActivity` per DEC-030), the listen-only flow (output independent of mic permission), the hard-mute override, and the 200 ms (0–500 ms) PTT release tail for word-boundary anti-clipping. |
 
 
 ## Baseline Candidate 0.9.9 Update