extract_content_tags now generates with grammar=None and parses the
response via a manual brace-counting JSON envelope extractor that
handles Qwen <think>...</think> prefixes and trailing stop tokens.
Five new unit tests. Bumps llama-cpp-2 to 0.1.146. Explicit
features=[] on tauri dependency (no-op).
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Replaces 22 production eprintln! sites with structured tracing events
across 8 files. Closes Area B of the post-prognosis residuals plan
(docs/superpowers/plans/2026-05-12-engine-slop-residuals.md).
Files touched (22 sites):
- crates/hotkey/src/linux.rs (2) — hotplug watcher degraded-mode warnings
- crates/ai-formatting/src/pipeline.rs (1) — LLM cleanup fallback warning
- src-tauri/src/commands/transcription.rs (1) — chunking dispatch info
- src-tauri/src/commands/diagnostics.rs (1) — crashes-dir setup warning
- src-tauri/src/commands/tasks.rs (1) — malformed feedback row warning
- src-tauri/src/commands/power.rs (3) — App Nap acquire/release/fail
- src-tauri/src/commands/models.rs (5) — Whisper warmup lifecycle
- src-tauri/src/commands/live.rs (8) — session start, chunk dispatch,
per-chunk delivery, inference errors, worker disconnects, listener
loss, status-channel cascade
Levels: error for unrecoverable failures (inference disconnect, panic,
status cascade), warn for recoverable degradation (LLM fallback,
malformed rows, App Nap fail, hotplug watcher fail), info for lifecycle
(session start, chunk processed, App Nap acquire/release, warmup
complete, chunking dispatch), debug for per-chunk noise (speech-gate
skip, chunk dispatch).
Two new dependencies and two new filter targets:
- tracing = "0.1" added to crates/hotkey and crates/ai-formatting
- Default EnvFilter in src-tauri/src/lib.rs::init_tracing extended with
magnotia_hotkey=info,magnotia_ai_formatting=info so the new targets
emit at the default level
Out of scope (intentional, left as-is):
- crates/mcp/src/main.rs — CLI binary, stderr is the log contract
(module docstring) so the JSON-RPC stdout stream stays clean
- crates/*/tests/*.rs and crates/core/examples/tuning_log_demo.rs —
test/example diagnostic output relies on --nocapture stdio semantics
Discovery during sweep (not fixed — separate follow-up): hotkey crate
has 6 existing log:: calls (log::error/warn/info/debug) but the
workspace builds tracing-subscriber without the tracing-log feature, so
those events are currently silent. Worth a follow-up to either add the
tracing-log bridge or migrate hotkey's existing log:: calls to
tracing::.
Verification:
- cargo fmt --all
- cargo check --workspace --all-targets — clean
- cargo test --workspace — 330+ tests, zero failures
- rg eprintln! src-tauri/src/commands/ crates/hotkey/src/ crates/ai-formatting/src/ — zero hits
Pre-existing working-tree churn in crates/llm/, src/lib/pages/,
src/lib/utils/saveMarkdown.ts and the untracked phase10a dogfood notes
deliberately left unstaged per Jake's instruction.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
The 2026-05-12 engine-slop pass removed runtime std::env::set_var mutation from src-tauri/src/lib.rs and replaced it with warn_if_x11_env_unset_on_wayland. With no launcher owning the contract, Linux Wayland dogfood was about to regress.
run.sh:
- now owns Linux launcher env defaults via case "$(uname -s)"
LIBCLANG_PATH=/usr/lib64/llvm21/lib64 (user-set wins)
WEBKIT_DISABLE_DMABUF_RENDERER=1 (always on Linux; user-set wins)
GDK_BACKEND=x11, WINIT_UNIX_BACKEND=x11 (Wayland only; user-set wins)
- 60s Vite readiness timeout
- detects early Vite exit (kill -0 + wait) instead of hanging forever
- trap installed before wait so Ctrl-C cleans up
- args forwarded: ./run.sh --release etc.
- non-exec final Tauri launch preserved so cleanup trap fires
package.json:
- "dev:tauri": "./run.sh" — canonical discoverable dev command
Docs:
- README, dev-setup, architecture-map runtime + launcher pages updated with the new contract; canonical command is npm run dev:tauri (./run.sh as direct equivalent)
- dev-launcher-and-scripts.md replaces the incorrect "kills process group" claim with honest "kills the spawned npm process"
- dev-setup.md path /CORBEL-Projects/magnotia → /CORBEL-Projects/transcription-app
- gpu-tuning/plan.md gets a superseded note rather than rewriting the original rationale
- engine-slop-residuals.md gains Area F (packaged-binary launcher contract) with honest wrapper-vs-.desktop trade-off documented
Verification: bash -n run.sh; shellcheck clean; cargo check -p magnotia green; npm pkg get 'scripts.dev:tauri' returns "./run.sh"; stale-ref sweep clean across living docs.
Sweep of registered-but-never-invoked commands surfaced by an audit
against the frontend invoke() call sites. Each was confirmed dead via
grep across src-tauri, src, and crates: no caller anywhere.
Removed commands:
- check_model, count_transcripts_command, get_profile_cmd,
install_update, list_feedback_examples_cmd (utility/CRUD shapes
never wired)
- save_audio, start_native_capture, stop_native_capture (native-capture
path superseded by the live transcription session)
- transcribe_pcm, transcribe_pcm_parakeet (PCM commands superseded by
live session; no frontend caller)
- close_preview_window (preview window is hidden via the
core:window:allow-hide capability, not the command)
Cascade in audio.rs (~430 lines removed):
- CaptureWorker, NativeCaptureState struct + impl, stop_worker,
append_recorded_chunk, MAX_NATIVE_CAPTURE_RETURN_SAMPLES,
persist_audio_samples
- The two cfg(test) tests that exercised stop_worker (the
recording_filename tests stay, supporting resolve_recording_path
which the live session uses)
Cascade elsewhere:
- FeedbackDto struct and its From<FeedbackRow> impl
- Stale storage imports in feedback.rs, profiles.rs, transcripts.rs
- tauri::Emitter import in transcription.rs
- app.manage(NativeCaptureState::new()) in lib.rs setup
generate_handler entries removed for all 11 commands. cargo check passes
cleanly with zero warnings; no tests reference any deleted symbols.
Net: 709 deletions, 17 insertions across 9 files.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
transcribe_pcm and transcribe_pcm_parakeet did not check that their
respective engines were loaded before clone+spawn_blocking. transcribe_file
already calls ensure_model_loaded; these now mirror that posture with a
friendly error when the engine is unloaded, matching what
extract_content_tags_cmd does.
decompose_and_store and extract_tasks_from_transcript_cmd ran multi-second
LLM inference inside spawn_blocking without the PowerAssertion guard that
cleanup_transcript_text_cmd and extract_content_tags_cmd already use. Both
now begin a guard so the macOS App-Nap inhibitor (and the planned
Linux/Windows equivalents per KI-02, KI-03) can pin the process for the
duration of the inference.
HANDOVER.md gets a status note clarifying it captures Phase 9 state. The
schema head referenced (v14) was the head at session time; current head
is v15 (`idx_transcripts_profile_created` composite index) per the
architecture map.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
PowerAssertion file-level doc previously claimed Linux logind and
Windows SetThreadExecutionState implementations in present tense.
Both are no-ops; the macOS path compiles but is unverified on
Apple Silicon (RB-08). Rewrite top doc to state present vs
planned posture and reference KNOWN-ISSUES.md.
Surfaced as tracked limitations:
- KI-01: macOS App Nap guard pending Apple Silicon verification
- KI-02: Linux power assertion is a no-op
- KI-03: Windows power assertion is a no-op
- KI-04: magnotia-cloud-providers crate not user-exposed in v0.1
(in-memory keystore needs OS keychain before any save-key UX)
README links to KNOWN-ISSUES.md from Status, Platform support
table, and Project documentation. Platform support table notes
adjusted per OS to reflect actual idle-inhibit posture.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Delete the local duplicate fn and libloading dependency from src-tauri;
import the canonical implementation from magnotia-core::hardware instead.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Bundled work from a low-end laptop (Ryzen 5 4650U / Vega 6 / Linux Mint
22.2 / X11) profiling pass.
perf: WEBKIT_DISABLE_DMABUF_RENDERER=1 default on all Linux
Previously only set on Wayland sessions. Empirically it's a
significant idle-cost win on integrated GPUs in either session type:
env-var matrix (release binary, 75s settle, 10s jiffies CPU sample)
showed magnotia idle CPU 12.30% → 2.80% of one core and idle GPU
17% → 10% on this hardware. Users can opt back in by exporting
WEBKIT_DISABLE_DMABUF_RENDERER=0. The Wayland-only XWayland
fallback (GDK_BACKEND=x11, WINIT_UNIX_BACKEND=x11) is unchanged.
fix: secondary-windows ACL — allow set-always-on-top
The float window's pin toggle calls setAlwaysOnTop() but the
secondary-windows capability didn't permit it, so the popout was
stuck always-on-top regardless of the pin state. Adds the
core:window:allow-set-always-on-top permission. Narrow scope.
fix: guard registerGlobalHotkey against non-main webviews
Cross-window settings sync via localStorage can re-fire the
$effect(() => settings.globalHotkey) callback inside popout webviews
where the main layout's registerGlobalHotkey is reachable. Adds an
early-return when the current window label is not "main", so the
popout doesn't trigger an ACL-denied register/unregister and the
user no longer sees a spurious "Hotkey not registered" toast when
popouts are open. Keeps the global-shortcut perm scoped to main.
build: pin @tauri-apps/api 2.10.1 + @tauri-apps/plugin-dialog 2.7.1
Match the Rust crate versions tauri-cli's version-mismatch check
enforces during release builds. Without this, `npm run tauri build`
exits 0 silently while emitting an Error and never producing
binaries.
test: add crates/transcription/tests/jfk_bench.rs
Reproducible RTF regression fixture. Env-gated on
MAGNOTIA_WHISPER_TEST_MODEL + MAGNOTIA_WHISPER_TEST_AUDIO so it
never runs in CI without setup. Loads the JFK WAV inline (no hound
dep), times model load + cold + warm transcribe, prints SUMMARY.
Baselines on this hardware:
--release --features whisper: cold RTF 0.054, warm 0.050, RSS 125 MB
--release --features whisper,whisper-vulkan: cold RTF 0.029, warm 0.028, RSS 125 MB
Vulkan on RADV/Vega 6 nearly halves transcription latency for
Whisper Tiny — useful baseline for Phase 10 hardware-recommendation
scoring.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Picks up the registry rename in the front-end and Tauri command layer:
- src/lib/types/app.ts: LlmModelIdStr now lists the four new ids
(qwen3_5_2b / qwen3_5_4b / qwen3_5_9b / qwen3_6_27b).
- src/lib/pages/SettingsPage.svelte: LLM_MODELS table rebuilt with
four tiers (Minimal / Standard / High / Maximum), matching subtitles
and download-size copy. selectedLlmModelId fallback, hardware-warning
thresholds, tier-availability check, and ensureRecommendedLlmTier
fallback all retargeted at the new ids. The Maximum tier surfaces a
64 GB / 24 GB warning so users with mid-range hardware see honest
expectations.
- src-tauri/src/commands/llm.rs and commands/tasks.rs: doc-comment
examples refreshed (Qwen3 4B → Qwen3.5 4B, Qwen3's tokenizer →
Qwen's tokenizer — the BPE family is shared).
- src/lib/stores/llmStatus.svelte.ts: chip-detail example updated.
cargo build --workspace clean. cargo test --workspace clean.
npx svelte-check reports one pre-existing error in vite.config.js
(unused @ts-expect-error directive, dates back to the original
scaffold commit 9926a42); not introduced here, out of scope to fix.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Replace all instances of the legacy product names "Kon" and "Corbie" with
"Magnotia" across user-facing copy, code identifiers, package names, bundle
ids, file paths, and documentation. Preserves the unrelated "konsole" (KDE
terminal) reference and the parent CORBEL company name.
- Renames 10 Rust crates (kon-* → magnotia-*) and the tauri binary
- Updates package.json, tauri.conf.json (productName + identifier)
- Renames CSS classes (kon-rh-* → magnotia-rh-*) and animations
- Renames brand and roadmap docs
- Regenerates Cargo.lock and package-lock.json
Verified: svelte-check passes; pure-rust crates compile under new names.
Two small Phase 1 follow-ups for the Android target:
1. tauri.conf.json: add `bundle.android.minSdkVersion: 24`. Android 7.0
is the floor — gives us Vulkan availability (for the eventual GPU
feature flag), AAudio for cpal, and is what Pixel-class hardware
tests against. Keeps the global `identifier` on `uk.co.corbel.kon`
for now; the Corbie rebrand sweep will land `corbel.technology.corbie`
as a single coherent commit.
2. src/lib/utils/runtime.ts: add `isAndroid()` and `isMobile()` helpers
alongside the existing `hasTauriRuntime()`. Both use UA sniffing —
sufficient for feature-gating UI, never for security decisions.
These are how the Svelte side will hide:
- hotkey config (no global hotkey API on Android)
- paste-mode picker (auto-paste maps to a copy-only flow)
- meeting auto-capture toggle (process list unavailable)
- multi-window buttons (open-viewer, open-float, etc.)
- system-tray-related affordances
Tauri 2 doesn't expose a synchronous platform-detection helper that
works during initial render, so UA sniffing is the pragmatic choice.
https://claude.ai/code/session_0189xUb6ie6t9qHkzatGZ9Rb
Phase 1 of the Android same-repo target plan: make the workspace
compilable for `aarch64-linux-android` (and the other NDK ABIs) by
removing the desktop-only crate dependencies and command bodies from the
Android build. After this commit, `tauri android init` followed by
`cargo tauri android build` is structurally unblocked — the remaining
work is the SDK/NDK toolchain (off-sandbox), the Svelte single-window
refactor, and the Phase 3 MVP feature surface.
What's gated under `cfg(not(target_os = "android"))`:
- src-tauri/Cargo.toml: `tauri = { features = ["tray-icon"] }` is now
declared in the desktop-only target block. The `global-shortcut`,
`window-state`, and `autostart` plugins join it — none of the three
support Android natively. The base `tauri = "2"` plus `dialog`,
`opener`, and `notification` plugins remain unconditional because they
do support Android.
- src-tauri/src/lib.rs: `mod tray` declaration, the matching
`tray::setup(app)` call, the close-to-tray `WindowEvent::CloseRequested`
handler, and the `.plugin(tauri_plugin_global_shortcut::*)` /
`_autostart` / `_window_state` chain are all desktop-only. The
builder is split with a single `#[cfg(not(target_os = "android"))]`
branch that adds the desktop plugins on top of the universal base.
- src-tauri/src/commands/tts.rs: `tts_speak` previously had three
`#[cfg(target_os = ...)]` branches but no fallback, so on Android the
`spawned` binding was unbound and the function failed to compile.
Mirrored the existing `paste.rs` not-implemented fallback. Same fix
for `list_voices_impl`. Frontend will hide the Read Page Aloud button
on Android via `isAndroid()`.
- src-tauri/src/commands/windows.rs: all four multi-window commands
(`open_task_window`, `open_preview_window`, `close_preview_window`,
`open_viewer_window`) get an Android stub that returns a clear
"Multi-window is not supported on Android" error. Tauri on Android
is single-Activity; the previously-secondary content (preview overlay,
transcript viewer, task float) will live as routes inside the main
window, gated by `isAndroid()` on the frontend.
What's *not* changed:
- Top-level `identifier` in tauri.conf.json stays `uk.co.corbel.kon`.
The Phase 10b Kon → Corbie rename sweep will land
`corbel.technology.corbie` as part of a coherent rebrand commit
rather than fragmenting the rename across this branch.
- `bundle.android.minSdkVersion: 24` added so a future
`tauri android init` knows to target Android 7.0+ (Vulkan available,
scoped storage starts at 29 — we'll surface scoped-storage paths
via Tauri's dialog plugin on Phase 3).
- `kon-hotkey` already exports a non-Linux stub; no changes needed.
- `commands/meeting.rs` still calls `process_watch::list_running_process_names()`
which compiles on Android but returns an empty list (SELinux blocks
/proc walk on API 24+). Frontend will hide the toggle on Android.
Verification: 91/91 tests still pass on the buildable-in-sandbox crates
(kon-storage 60, kon-core 16, kon-mcp 9, kon-hotkey 4, kon-cloud-providers
2). svelte-check 0/0 across 3957 files. The src-tauri crate itself can't
be compiled in this sandbox (no webkit2gtk); CI's desktop builders will
exercise the desktop branch, and Jake's Android-equipped dev box will
exercise the Android branch via `tauri android init`.
https://claude.ai/code/session_0189xUb6ie6t9qHkzatGZ9Rb
`emit_live_result` already detected a lost result_channel listener: it
sent a one-shot status warning and from then on short-circuited future
result sends. But if the status_channel listener was also gone — which
is what happens when the user closes the main window without calling
stop_live_transcription_session — the worker kept polling inflight
inference every 10 ms forever, holding a model loaded on the GPU and
keeping the WAV writer file handle open until the process exited.
When the warning send to status_channel also returns Err, the entire
frontend channel pair is dead. Self-assert stop_flag from inside
emit_live_result so the worker drains and exits cleanly. Existing user-
initiated stop semantics are unchanged.
- Threaded `stop_flag: &Arc<AtomicBool>` through `emit_live_result` and
the free `poll_inference` (instance method already had access via
`self.stop_flag`).
- Existing `result_listener_loss_is_warned_once_*` test updated to pass
a stop_flag and assert it stays false when only result_channel fails.
- New test `dead_result_and_status_channels_self_assert_stop_flag` proves
the self-stop fires when both channels Err.
(src-tauri doesn't build in the audit sandbox — needs webkit2gtk; CI
cross-platform compiles it.)
https://claude.ai/code/session_0189xUb6ie6t9qHkzatGZ9Rb
The error_log table had no retention policy: every backend error was
appended forever, so across months of dogfooding it grew unbounded. That
silently bloats the diagnostic-bundle export and slows the
list_recent_errors query the Settings → About panel runs.
- New `kon_storage::prune_error_log(pool, keep_days)` does a single
`DELETE FROM error_log WHERE timestamp < datetime('now', '-Nd days')`
and returns the row count removed.
- src-tauri/src/lib.rs runs it once during setup() with a const
ERROR_LOG_RETENTION_DAYS = 90. Failure is logged to stderr but does not
block startup — a prune that fails is strictly less important than the
app coming up.
- Test: insert three rows at now / -30d / -200d, verify a 90-day prune
removes only the oldest, and a subsequent 14-day prune removes the
-30d row. Storage suite at 60/60.
https://claude.ai/code/session_0189xUb6ie6t9qHkzatGZ9Rb
`detect_meeting_processes` is called every 15 s when meeting-auto-capture
is enabled. The previous `list_running_process_names` allocated a fresh
`sysinfo::System` per call and walked /proc cold; on a busy host
(~300 processes) that's ~50–100 ms of work, every poll, forever.
Add `kon_core::process_watch::ProcessLister`, a thin wrapper around a
long-lived `System` whose process table is refreshed in place. The Tauri
host holds one behind a `Mutex<ProcessLister>` in a new `MeetingState`
managed at app setup. The free `list_running_process_names` is kept as a
convenience that constructs a fresh `ProcessLister` per call — its only
remaining caller is the existing smoke test.
- ProcessLister + Default in crates/core/src/process_watch.rs.
- MeetingState in src-tauri/src/commands/meeting.rs; the command takes
it via `tauri::State` and locks for the duration of the snapshot.
- src-tauri/src/lib.rs registers MeetingState alongside the other
managed states.
https://claude.ai/code/session_0189xUb6ie6t9qHkzatGZ9Rb
Adds llm_tags TEXT NOT NULL DEFAULT '' to the transcripts table via
new migration v14. SELECT statements + TranscriptRow + transcript_row_from
now carry the column. update_transcript_meta gains a sixth Option for
llm_tags following the existing COALESCE pattern; an
#[allow(too_many_arguments)] keeps clippy happy without inverting the
signature into a struct that would just shift the indirection.
The Tauri-side TranscriptDto + UpdateTranscriptMetaRequest + the
update_transcript_meta_cmd command pass llm_tags through unchanged.
Pre-existing manualTags persistence path now has a sibling for
llmTags ready for the frontend to call.
Phase 8 brittle test fix included: list_recent_completions_uses_local_day_boundary
was anchoring its "-2 days" UTC offset against the local-day spine,
which drifted across UTC midnight. Anchored to the local date directly
so it matches the spine regardless of clock.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Bridges LlmEngine::extract_content_tags to the frontend with the same
spawn_blocking + PowerAssertion guard the cleanup_text command uses.
Returns a ContentTags object serialised to camelCase JSON. Errors
surface as readable strings so the frontend toast shows actionable
text on the rare grammar-bypass path.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Thin UTF-8 writer used by the new save-dialog path. Caller owns path
safety; the source path is always OS-dialog-provided. Two unit tests:
roundtrips a small UTF-8 string with non-ASCII chars and asserts a
nonexistent parent path returns an actionable error.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Thin wrapper over kon_storage::list_recent_completions, parameterised
by day count. Serialises to camelCase JSON (day, count).
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Small if-then automation layer. Rules persist in SQLite; the runner
lives on the frontend and binds to the Phase 6 event bus so the rule
pipeline reuses the same delivery primitives (timer events, TTS,
Tasks navigation).
Storage:
- Migration v12 adds implementation_rules (id, enabled, trigger_kind,
trigger_value, actions_json, last_fired_key, created_at,
updated_at) with enabled+trigger_kind index for the runner's hot
path.
- CRUD helpers: insert / list / get / set-enabled / mark-fired /
delete, plus a round-trip test.
Commands (all main-window-guarded via ensure_main_window):
- list_implementation_rules
- create_implementation_rule — validates HH:MM, checks the target
task exists at save time for surface-task actions, caps
speak-line at 240 chars, pins v1 timers to 5 minutes.
- set_implementation_rule_enabled
- mark_implementation_rule_fired — main-thread idempotency shim so
the runner can atomically claim a fire.
- delete_implementation_rule
Runner (implementationIntentions.svelte.ts):
- Subscribes to kon:task-completed and kon:morning-triage-finished
(MorningTriageModal now emits on all three exit paths — empty,
skipped, picked — so skip counts as finishing).
- 30 s poll for time-of-day rules, plus an immediate check on startup
so a rule whose time has already passed today catches up once.
- Idempotency via last_fired_key composed as YYYY-MM-DD@HH:MM for
time rules; new time rules whose HH:MM has already passed today
are pre-seeded so they don't fire retroactively on save.
- Rules are paused when Nudges "Mute for now" is on — a hard mute
stops all rule delivery in addition to OS notifications.
- Stale-task safety: if a surface-task action's target has been
deleted, the runner opens Tasks and warns clearly rather than
pretending to surface something that's gone.
Editor (ImplementationRulesEditor.svelte):
- Lives in Settings under a new "If-then rules" accordion section.
- `If` picker: time of day (with time input), a task completes,
morning triage finishes.
- `Then` composer: optional surface (inbox / today / all tasks /
specific task), optional 5-min timer, optional speak-aloud line.
- Saved rules list with enable toggle + delete.
Rules table integration for Phase 10b rename sweep: add
implementation_rules to the kon.db → corbie.db migration shim when
that phase lands.
Gates: fmt, clippy -D warnings, cargo test 265/0, svelte-check
0/0, npm build green. Pre-existing Vite chunk warning on sounds.ts
is unrelated to Phase 7.
Frontend-owned nudge bus that consumes in-app signals Corbie already
produces, applies suppression, and fans out to OS notification + an
optional TTS read-aloud. OS-wide keyboard/window activity detection
stays deferred per the revised roadmap — the plan before rewrite
would have been brittle on Wayland, permission-heavy on macOS, and
low-quality everywhere.
Triggers (v1, all in-app signals):
- inactivity_with_active_timer — timer running, window blurred ≥ 90 s,
at least 60 s into the timer.
- pending_morning_triage — ritual enabled, past 10:00 local, last
shown ≠ today. Polls every 5 min while focused.
- micro_step_idle — micro-step decomposition created, no child step
or parent task completed within 15 min.
Suppression:
- Respects nudgesEnabled + nudgesMuted.
- No nudge while the app has focus (document.hasFocus).
- Hard cap 3 per rolling hour.
- Permission requested via @tauri-apps/plugin-notification on first
delivery; denial is silently respected.
Rust side:
- tauri-plugin-notification registered + ACL entries on the main-
window capability only (secondary windows can't fire nudges).
- commands::nudges::deliver_nudge — thin wrapper, security-guarded
via ensure_main_window, delegates to the plugin. No DB writes —
the roadmap's nudges-audit table is deferred until a concrete need
emerges.
Frontend glue:
- nudgeBus.svelte.ts — subscribes to window events, applies
suppression, calls deliver_nudge (+ tts_speak when speakAloud is
on).
- kon:task-completed now dispatched on complete_task_cmd success.
- kon:microstep-generated + kon:step-completed dispatched from
MicroSteps so the idle trigger can clear itself on any engagement.
- kon:focus-timer-cancelled added to focusTimer so the bus can reset
its inactivity state on cancel, not only on natural completion.
- nudgeBus started from +layout.svelte onMount, stopped on destroy.
Settings:
- New "Nudges" section with three toggles: Enable nudges, Mute for
now (separate so a hard mute doesn't lose preferences),
Speak nudges aloud (reuses Phase 4 TTS).
- All default OFF. No first-run prompt — nudges are a Settings-found
feature rather than a walkthrough step.
Out of scope per the revised Phase 6 spec: OS-wide keyboard/window
hooks, biometric signals, custom trigger editor (Phase 7), notification
sound (platform variance too high for Layer-1 — revisit in Phase 9
polish with a bundled .wav).
Gates: fmt clean, clippy -D warnings clean, cargo test 262/0 (4
existing + 262 current), npm run check 0/0, npm run build green.
Three opt-in rituals, all default OFF. Research-anchored (Barkley's
point-of-performance, Sweller cognitive-load theory, Newport shutdown
ritual, Gollwitzer implementation intentions, Thaler/Sunstein nudge
with informed consent for the ADHD audience).
Morning triage: modal gated on ritualsMorning toggle, configurable
trigger time (default 08:00 to respect ADHD sleep inertia rather than
the spec's 06:00), "pick up to three for today" with a gentle swap
message on the fourth attempt. Skip sets last-shown-today so it never
re-prompts the same calendar day. last-shown persists via kon_storage.
Evening wind-down: dedicated page, user-triggered only (tray menu +
Settings button). Mechanical closure + physical reset + intentional
cue — the whole Newport template. Open loops are read-only reflection;
Tasks page owns transactions. Copy is additive throughout: "you
finished X today", never "you didn't finish Y".
Autostart: tauri-plugin-autostart registered (LaunchAgent on macOS,
.desktop on Linux, registry Run on Windows). No bespoke Rust commands
— frontend calls the plugin's invoke-handlers directly. Toggle in
Settings is one-way (click → OS call → state update) to avoid the UI
lying during the round-trip. First-run presents a forced-choice prompt
for all three options, with "skip all" escape hatches per step.
Copy audit against RSD literature: no "overdue", "failed", or
day-to-day comparison framing anywhere in ritual surfaces.
Post-v0.1 ideas captured in the roadmap: calendar integration
(read-only ICS as interim, cloud sync parked) and right-click-to-task
(in-app simple, system-wide a separate phase).
Platform-dispatched TTS (spd-say + espeak-ng fallback on Linux, say on
macOS, PowerShell System.Speech on Windows) with a shared SpeakerButton
component. Tap to speak, tap again to stop; only one button speaks at
a time so two surfaces don't talk over each other. Text always travels
via argv (or a PowerShell here-string delivered through -EncodedCommand
on Windows) so user content never enters a shell string.
Mount points: DictationPage transcript footer, transcript viewer header,
per-step in MicroSteps. Settings gains a "Read aloud" accordion with
voice picker (lazy-loaded from the OS synth), rate slider 0.5-2.0x,
and a British-English test utterance.
Rust tests cover rate mapping, NaN handling, and Windows here-string
terminator safety. No pause/resume, no SSML, no cloud voices — that
stays out of scope per the Layer-1 roadmap.
Closes Phase 3 of the 2026-04-23 feature-complete roadmap. Incorporates
the Codex plan-review fixes from this session: profile-free index, tri-
state update command, and de-prioritise-not-hide semantics.
Storage (kon-storage):
- Migration v11 adds `energy TEXT` to `tasks` with a CHECK constraint on
`high | medium | brain_dead | NULL`. Index `(energy, created_at DESC)`
— deliberately not per-profile because the tasks table carries no
profile_id column yet (tracked as a separate gap in HANDOVER).
- `TaskRow.energy: Option<String>` plus `task_row_from` read.
- `insert_task` signature grows by one optional arg (`energy`). Allowed
`too_many_arguments` with a rationale comment — the positional shape
matches the column order and flipping to a params struct would have
rippled through every caller for cosmetic benefit only.
- New `set_task_energy(pool, id, Option<&str>) -> TaskRow`. Lives as its
own function because `update_task` uses COALESCE to let `None` mean
"preserve" — which would make clearing the tag impossible.
- Two new tests: round-trip including explicit NULL clear, and CHECK
constraint rejection of unknown values.
- Tests updated for the v10 → v11 version bump.
Tauri (src-tauri):
- `TaskDto.energy`. `CreateTaskRequest.energy` (optional). Inline
validation against the allowed set before hitting the DB, so frontend
bugs surface as friendly errors instead of CHECK-constraint failures.
- New `set_task_energy_cmd` command mirroring the storage tri-state API.
Frontend (svelte):
- `EnergyLevel` type added to `types/app.ts`. `TaskDto`, `TaskEntry`, and
`TaskDraft` grow an `energy` field.
- `SettingsState.currentEnergy` (persisted) + `matchMyEnergy` (persisted
toggle). Defaults: null + false — no surface change until user opts in.
- `setTaskEnergy(id, EnergyLevel | null)` action on the task store.
Calls the dedicated Tauri command, updates local state, broadcasts to
sibling windows.
- `EnergyChip.svelte` — new component. Cycles unset → High → Medium →
Brain-Dead → unset on click. Colour tokens: accent / warning /
text-tertiary (deliberately not danger-red for Brain-Dead — the brief
is explicit that this state must not feel pathologised).
- Chip rendered on every task row in TasksPage and every row in
WipTaskList. Hidden-until-hover when energy is unset so untagged rows
stay calm; always visible once tagged because the colour is the signal.
- Tasks page header gains a "I feel" segmented control and a
"Match my energy" toggle. When both are active, matching tasks sort
to the top — unset tasks are treated as Medium-equivalent. Nothing is
ever hidden; this is a de-prioritisation, not a filter.
Deferred / out of scope:
- LLM-driven surfacing (brief says "The AI surfaces...") — deterministic
client-side sort is v1; LLM layer is a later phase.
- tasks.profile_id + per-profile energy sort — separate migration.
All green: cargo build + 251 tests + clippy -D warnings (0 warnings)
+ fmt + svelte-check (0/0) + npm run build.
Independent review surfaced three majors and one nit. All actioned.
MAJOR 1 — profile scoping:
`decompose_and_store` and `extract_tasks_from_transcript_cmd` now
accept an optional `profile_id` (wired from `profilesStore.activeProfileId`
in MicroSteps.svelte and DictationPage.svelte), and thread it into the
feedback-retrieval query so per-profile decomposition styles do not
leak into each other. `record_feedback` gets the same treatment.
MAJOR 2 — prompt-budget regression on long inputs:
New `trim_to_budget` helper + `FEW_SHOT_CHAR_BUDGET = 2000` char cap
in `src-tauri/src/commands/tasks.rs`. Retrieval still pulls up to 5
rows but they are char-counted and truncated against the budget
before being sent to the LLM. Char cost matches the `Input: ...\n
Good output: ...` render path so the budget maps cleanly to ~570
Qwen3 tokens, well inside the 8192-context reserve after the 512-
or 768-token response allocation. Oldest-first drop order (iteration
stops at cost exceeded) preserves the most recent correction which
is the one carrying the user's live preference.
MAJOR 3 — inline edit stale-rollback race:
`saveEdit` in MicroSteps.svelte now stamps a monotonic per-step
`saveToken`. Each edit bumps the token; on failure the rollback
only fires if `saveToken[step.id] === myToken`, so a slow-failing
first save can no longer overwrite a faster successful second save.
NIT — retrieval ordering stability:
`list_feedback_examples` ORDER BY now `created_at DESC, id DESC`.
SQLite timestamp precision is one second; without the secondary
key, bursty feedback within the same second would select
non-deterministically.
Also: malformed `context_json` now warns via eprintln! rather than
disappearing silently — Codex minor.
All green: cargo build + 249 tests + clippy -D warnings + fmt
+ svelte-check (0/0) + npm run build.
Closes the human-in-the-loop gap from docs/brief/feature-set.md and
Phase 2 of the 2026-04-23 feature-complete roadmap.
Storage (kon-storage):
- Migration v10 adds the `feedback` table: (target_type, target_id,
rating, original_text, corrected_text, context_json, profile_id,
created_at) with CHECK constraints on target_type and rating, plus
indexes on (target_type, rating, created_at DESC) for prompt-time
retrieval and (profile_id, target_type, created_at DESC) for
per-profile scoping.
- New public API: `FeedbackTargetType`, `RecordFeedbackParams`,
`FeedbackRow`, `record_feedback`, `list_feedback_examples`.
- Tests updated — the RB-02 rollback regression now discovers the
real max version at runtime instead of hard-coding v10 for its
poison migration.
LLM (kon-llm):
- `prompts::FeedbackExample` — local shape for few-shot exemplars so
kon-llm stays independent of kon-storage.
- `prompts::build_conditioned_system_prompt` — appends a "here is
the style this user prefers" block to the base system prompt
when examples are available; returns the base prompt unchanged
when empty, so new users and early sessions see generic output.
- `LlmEngine::decompose_task_with_feedback` and
`LlmEngine::extract_tasks_with_feedback` thread examples through
to the builder. The old one-arg variants are preserved and now
call through with an empty slice.
- 4 unit tests covering empty, empty-input-skip, correction-wins,
and thumbs-up-only fallback.
Tauri (src-tauri):
- New commands::feedback module: `record_feedback`,
`list_feedback_examples_cmd`.
- `decompose_and_store` and `extract_tasks_from_transcript_cmd`
now fetch the last 5 positive/neutral feedback rows for their
target type and pass them through to the LLM, wiring the
learning loop end-to-end.
- Shared `to_llm_examples` helper parses the `context_json.input`
field (where the recorder stashes the parent task text / transcript
chunk) back into the exemplar shape.
Frontend (MicroSteps.svelte):
- Thumbs-up and thumbs-down buttons on every micro-step row.
Hover-revealed; the vote recolours the icon; clicking again
clears the local highlight (the row itself stays in the audit
trail).
- Pencil icon + double-click to edit step text. Save flows through
update_task_cmd for persistence and records a correction feedback
row with (original_text, corrected_text) — the highest-value
training signal.
- Parent task text is captured in context_json.input at record time
so the prompt builder can reconstruct the (input, preferred-output)
pair on subsequent decompositions.
- Feedback capture is best-effort — a record_feedback failure never
interrupts the primary action.
What's deferred to a later phase:
- Thumbs + corrections on extracted tasks (same pipeline, different
surface — probably TasksPage after the AI-extraction path)
- Thumbs on transcript cleanup output
- Semantic retrieval over the feedback corpus (once there is enough
data to justify embedding infrastructure; the storage shape is
already ready for it)
Auto-applied cargo clippy --fix across 11 files — needless return,
unnecessary cast, map_or simplification, repeat().take() → repeat_n(),
iter().any() → contains(), manual char comparison, lifetime elision,
push_str single-char, reference immediately dereferenced.
Also fixed three lints on file_storage.rs manually: two doc-list-item
overindentations, plus the same needless-return. Baseline main was
not clippy-clean with -D warnings before; after this pass one
needless_range_loop warning remains (live.rs:1089) that clippy's
suggested rewrite would make less readable — left for a dedicated
refactor session.
Build + workspace tests remain green (245 passing, 0 failing, 1
ignored).
The accumulator task was fire-and-forget — `tokio::spawn` without
retaining the JoinHandle. `stop_native_capture` sent a stop signal,
slept 50ms, and returned; the worker could still be running its
final flush and appending to `all_samples` when the next
`start_native_capture` cleared it. Rapid start→stop→start could
leak tail samples from one session into another.
Replace `NativeCaptureState.stop_tx` with `worker:
AsyncMutex<Option<CaptureWorker>>`, where CaptureWorker owns both
the stop sender and the spawned task's JoinHandle. New helper
`stop_worker(worker)` sends stop, drops the sender, and `.await`s
the join. Both the prior-worker tear-down in `start_native_capture`
and `stop_native_capture` itself go through the helper, so the
worker is always fully terminated before any downstream read or
next-session cleanup.
AsyncMutex (not std::sync::Mutex) because the stop path awaits
while holding the lock. Also drops the 50ms sleep from
stop_native_capture — the join is an exact barrier.
Two regression tests:
- stop_worker_awaits_full_termination_no_writes_after_join:
synthetic worker with an atomic counter and a flush marker.
After stop_worker the flush must have run and no further
writes may appear.
- stop_worker_is_idempotent_on_a_worker_that_has_already_exited:
tasks that stop themselves must still join cleanly.
A full cpal-backed start→stop→start integration test is not
feasible in Linux CI without an audio device. The component tests
cover the invariant the real flow depends on.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
get_runtime_capabilities was returning `accelerators = ["cpu",
"vulkan"]` and `whisper.supports_gpu = true` regardless of build
config or runtime state. On a macOS build it falsely advertised
"vulkan" (the backend actually resolves via MoltenVK as Metal); on a
whisper-disabled build it claimed GPU support for an engine that
hadn't been linked.
Added `compose_accelerators(whisper_enabled, loader_available, target)`
— a pure helper that always emits "cpu" first and appends the
platform-appropriate GPU name only when whisper is compiled in AND the
Vulkan loader resolves. `supported_accelerators()` wraps it with the
live `cfg!(feature = "whisper")`, loader probe, and target OS.
`get_runtime_capabilities` now calls `supported_accelerators()` and
sets `whisper.supports_gpu = cfg!(feature = "whisper")`. Parakeet
stays CPU-only.
Five tests in `commands::models::tests` cover the permutation matrix:
whisper on/off, loader present/missing, macOS vs other. Both feature
configurations (`--features whisper` and `--no-default-features`)
build and pass tests.
Macos Metal-loader resolution on real hardware stays on the
ship-gate checklist — the detection logic is verifiable from Linux
but runtime behaviour is not.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
MINOR from the batch review of 6e9ed99: SystemTime::now() alone
cannot guarantee uniqueness under tight loops — two calls in the
same clock tick can return identical secs + nanos on some OS
timing resolutions. The filename reduction from "every second"
to "every nanosecond" addresses the flagged bug but leaves a
theoretical gap.
Adds a process-lifetime AtomicU64 counter, zero-padded to 4 digits,
as the third filename component. New shape:
kon-<secs>-<nanos_in_sec>-<counter>.wav
e.g. kon-1776828000-123456789-0000.wav
Across process restarts the counter resets to 0, but the wall-clock
secs/nanos have advanced — no cross-launch collisions possible.
Within a single process, the counter guarantees uniqueness regardless
of clock behaviour.
Test strengthened from ">=32 of 64 unique" (probabilistic) to
"1024 of 1024 unique" (absolute).
2026-04-22 review MINORs and NITs:
- crates/core/src/providers.rs: delete entire module. SpeechToText /
TextProcessor / ProviderRegistry were forward-looking traits that
never got wired — the Transcriber trait in kon-transcription
(A.2 #13) has since superseded SpeechToText, and the Registry
pattern was redundant against LocalEngine. Keeping them as dead
public surface signalled future direction that is no longer
accurate.
- crates/core/src/types.rs: delete TranscriptMetadata. Forward-
looking struct with an unfulfilled TODO; storage has evolved
independently through v7/v8 migrations without adopting it.
- crates/ai-formatting/src/llm_client.rs: remove #[allow(dead_code)]
from CLEANUP_PROMPT and format_dictionary_suffix. Both are
actively called; the suppressions would hide future genuine
dead-code warnings in this regression-sensitive prompt file.
- src-tauri/src/commands/live.rs: remove #[allow(dead_code)] from
LiveStatusMessage. Every variant (Warning, Overload, Error,
Finished) is constructed in the module today.
- README.md: update kon-transcription row from "SpeechToText
trait" to "Transcriber trait" and mention the new streaming/
module.
Workspace test gate green (225 lib tests across all crates).
MINOR from the 2026-04-22 review (build.rs:47-64): the guard used
strip_prefix("connect-src") to find the directive, which would also
match hypothetical future directives like connect-src-elem and
validate the wrong allow-list.
Switches to split_once(char::is_whitespace) on each directive and
requires the first token to equal "connect-src" exactly. Robust
against prefix collisions with any future CSP3 directive
additions.
MAJOR from the 2026-04-22 review (audio.rs:236-257): filenames were
derived from SystemTime::now().as_secs(), so two recordings started
within the same second resolved to the same path — possible overwrite
or merge.
Extracts the filename generation into a private helper and appends
the sub-second nanosecond component. Format is now
`kon-<secs>-<nanos_in_sec>.wav`, e.g. `kon-1776828000-123456789.wav`,
which stays human-readable, sortable by timestamp, and effectively
uncollidable under any realistic live-capture pattern.
Two tests cover the regression:
- recording_filenames_are_unique_across_rapid_calls: 64 tight-loop
calls must produce at least 32 unique names.
- recording_filename_has_expected_shape: prefix/suffix plus the
zero-padded 9-digit nanos component.
MAJOR from the 2026-04-22 review (paste.rs:181-217): unlike
paste_text, the replace flow did not snapshot the user's clipboard
before writing the raw transcript. A successful revert left the raw
transcript on the clipboard and destroyed whatever the user had
copied before invoking it — inconsistent with paste_text's Handy-#921
contract (brief item #3).
Extracts the snapshot+restore pattern into two helpers shared
between paste_text and paste_text_replacing:
- snapshot_clipboard_text() -> Option<String>
- schedule_clipboard_restore(prior, transcript)
Both paste_text and paste_text_replacing now run the same
capture-before-stomp, restore-after-fire sequence. The restore-
guard (only restore if clipboard still holds the transcript we set)
is shared too, so a user who copies something new within the 300 ms
window is still respected.
DRY: removes ~15 lines of inline clipboard bookkeeping from
paste_text while making the replace flow identical.
Review feedback: reported_audio_path was cached at writer-open time
and survived a mid-session writer drop. If an append error cleared
wav_writer (or hound's Drop ran instead of finalize), the Summary
still reported the path — risking StopLiveTranscriptionResponse
pointing at a file whose header did not reflect its data chunk.
Moves the decision to end-of-session. wav_writer.take() is inspected:
- writer present + finalize Ok → report the path,
- writer present + finalize Err → None, emit a Warning,
- writer absent (mid-session drop, or never opened) → None.
StopLiveTranscriptionResponse.audio_path now means "the recording is
known-good" rather than "a recording was attempted". Users can still
recover partial files via filesystem if needed; the warning toasts
already emitted by append_resampled_audio on write failure surface
that path implicitly.
Review feedback: src-tauri/src/commands/models.rs was still naming
load_whisper unconditionally, so a --no-default-features workspace
build (kon-transcription without the whisper feature) would have
compiled the transcription crate cleanly but failed in the kon crate
as soon as it tried to resolve the load_whisper symbol.
Adds a matching [features] section to src-tauri/Cargo.toml:
- default = ["whisper"]
- whisper = ["kon-transcription/whisper"]
and declares the kon-transcription dep with default-features = false so
the feature actually propagates rather than being forced-on by the
child crate's default. Cfg-gates the load_whisper import in models.rs
and adds a companion match arm that returns a runtime error when the
feature is off, keeping the API shape intact.
Verified with both cargo build -p kon and cargo build -p kon
--no-default-features.
The Vec<f32> in-memory accumulator on run_live_session had three
failure modes: (a) a crash during transcription took the recording
with it, (b) RAM grew linearly with session length, (c) OOM killed
the capture thread silently.
New kon_audio::WavWriter wraps hound::WavWriter<BufWriter<File>> with
an append-friendly API and a 500 ms-granularity header flush. On any
abort after a flush the on-disk file is a valid, playable WAV. Unit
test (brief item #19 acceptance) simulates the abort with
std::mem::forget and asserts the pre-flush samples are recoverable.
Live capture now:
- resolves the destination path at start_live_transcription_session
time via a new resolve_recording_path helper extracted from
persist_audio_samples,
- opens a WavWriter before any samples arrive, sample rate taken from
LocalEngine::capabilities() (#13 wiring) with 16 kHz fallback,
- feeds the resampler output through WavWriter::append inside
append_resampled_audio — drops the writer with a user-visible
warning if a write fails mid-session,
- calls flush() at stop (after resampler tail), finalise() on clean
exit, and drops-to-last-flushed state on abort.
LiveSessionSummary.audio_samples → audio_path: the path is already
written by the time stop_live_transcription_session runs; no
post-session write step remains for live capture.
persist_audio_samples is kept for the offline save_audio command.
New crates/transcription/src/transcriber.rs defines a Transcriber
trait (Send supertrait for spawn_blocking travel) with
TranscriberCapabilities (sample_rate, channels, supports_initial_prompt).
TranscriberCapabilities.sample_rate is load-bearing for the upcoming
progressive WAV writer (#19).
Concrete impls:
- SpeechModelAdapter wraps Box<dyn transcribe_rs::SpeechModel + Send>
for Parakeet (and any future transcribe-rs-backed engine).
- WhisperRsBackend moves its transcribe_sync body into the impl,
widening the signature from &self to &mut self so per-call
WhisperState can be created cleanly through the trait object.
LocalEngine now holds Box<dyn Transcriber + Send>; dispatch in
transcribe_sync collapses from a match to a direct call. Adds
LocalEngine::capabilities() for the WAV-writer.
Cargo feature flag "whisper" (default on) makes whisper-rs, num_cpus,
and the whole whisper_rs_backend module optional. cargo check
--no-default-features -p kon-transcription now builds without pulling
whisper-rs-sys — the escape hatch brief item #6 / #13 called for on
Windows / non-AVX2 / cloud-only builds. load_whisper is cfg-gated
behind the same feature.
src-tauri/src/commands/models.rs load_model_from_disk returns
Box<dyn Transcriber + Send> instead of SpeechBackend; caller chain
(ensure_model_loaded, prewarm_default_model) is unchanged.
transcriber_trait_is_object_safe test lands alongside the trait as a
compile-time witness against future Self-returning / generic-method
additions.
Review feedback: the original guard substring-searched the whole file
after the first "csp" token, which (a) false-passes if any unrelated
JSON value elsewhere in the config happens to contain a localhost URL
and (b) false-fails if the CSP is ever re-serialised with escaped
forward slashes.
Switches to serde_json + a /app/security/csp pointer lookup, then
splits the CSP on ';', finds the connect-src directive, tokenises its
allow-list on whitespace, and requires an exact match for both
http://127.0.0.1:* and ws://127.0.0.1:*. The error now also includes
the current connect-src value so a developer who breaks it can see
exactly what needs restoring.
Pins the connect-src CSP entries for http://127.0.0.1:* and
ws://127.0.0.1:* at build time. If a future edit to tauri.conf.json
strips the local-LLM permit, the kon crate fails to build rather than
shipping a binary whose webview fetch() silently 404s with an opaque
scope error (Vibe #438 / #487).
Closes item #2 of docs/whisper-ecosystem/brief.md — CSP widening itself
landed in an earlier commit; this is the regression-proofing the plan
calls for.
Two new Settings → AI knobs that compose cleanly with what already
shipped (aiTier, LLM model, translator prompt framing).
**B.1 #15 — Named cleanup presets.** LlmPromptPreset enum
(Default / Email / Notes / Code) appends a short context hint onto
the CLEANUP_PROMPT just before generation. Presets shape tone and
structure ("email paragraph", "bulleted meeting notes", "preserve
technical terms") without licensing the content-editing the
translator-not-editor framing forbids. cleanup_transcript_text_cmd
now takes `preset: Option<String>` which runs through the new
LlmPromptPreset::parse (normalises aliases like "meeting-notes",
collapses unknown values to Default).
**A.1 #28 — Sequential-GPU guard.** New LocalEngine::unload drops
the backend + model_id so a subsequent load actually reclaims VRAM.
load_llm_model, load_model, and load_parakeet_model Tauri commands
grow an optional `concurrent: bool` argument. When concurrent is
Some(false), loading LLM first unloads whisper+parakeet, and vice
versa — prevents VRAM OOM on tight-VRAM setups. Default is the
previous parallel behaviour so nothing changes for multi-GB cards.
Transcribe-in-progress paths (transcribe_pcm, transcribe_file, live)
pass None, so mid-dictation model loads don't accidentally tear
down the LLM.
Settings UI (AI section):
- Cleanup preset segmented button + descriptive copy for each option.
- GPU concurrency segmented button with explicit trade-off text
("faster transitions vs fits in tight VRAM").
Frontend wiring:
- settings.llmPromptPreset flows from DictationPage's
cleanupTranscriptIfEnabled into the Tauri command.
- settings.aiGpuConcurrency flows from both DictationPage (auto-load
on record) and SettingsPage (manual load/unload buttons) as
`concurrent: "parallel" === true` to the load commands.
Tests: three new preset cases in crates/ai-formatting/src/llm_client.rs
(parse aliases, suffix non-empty for non-default, default suffix
empty). All 139 existing lib tests still pass.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
The brief's pain point is opaque load failures: llama-cpp-2's errors
bubble up as raw C++ strings ("cudaMalloc failed: out of memory",
"invalid gguf magic"). A user seeing that has no path to recovery.
New backend command test_llm_model runs a staged diagnostic:
1. Model not downloaded → `not-downloaded` + download hint.
2. File size ≤90% of expected → `incomplete` (stalled download)
+ re-download hint. Matters because llama-cpp-2 can segfault
on truncated GGUF rather than returning cleanly.
3. Requested model already loaded → `ready`, no side effects.
4. Otherwise attempt a real load. On failure, classify_llm_load_error
maps the raw string to one of:
- load-failed-vram (OOM / cudaMalloc / allocation)
- load-failed-corrupt (GGUF magic / unsupported format)
- load-failed-permission (permission denied / access denied)
- load-failed-other (catch-all)
Each category has a prewritten actionable hint pointing at the
specific Settings surface (tier picker, re-download, file perms).
classify_llm_load_error is pure-string and unit-tested — 8 cases
covering the main categories plus edge cases (OOM alias, Windows
"Access is denied", unknown errors). Ordered narrow-to-broad so
overlap doesn't misclassify.
Settings UI gets a "Test" button in the AI section's action row,
visible whenever the model is downloaded (both downloaded-idle and
loaded states). Shows inline hint below the status line when the
test surfaces one. Refreshes both local and global LLM status after
the test since a successful test implicitly loads the model.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Kon's ideology rule: raw Whisper output is the source of truth; LLM
cleanup is additive, never destructive. The preview overlay already
tracks both rawText and finalText across the listening → live →
cleanup → final phases — but until now the user had no one-click path
from final to raw if cleanup changed their meaning.
Frontend: a context-aware "Use raw" / "Copy raw" button appears in
the preview overlay's final phase, only when rawText and finalText
actually differ (Raw format mode or LLM-off leaves the button hidden).
Two behaviours depending on how the transcript reached the target:
- settings.autoPaste = true → invoke paste_text_replacing, which
sends the platform's undo keystroke to the focused app,
waits UNDO_PASTE_GAP_MS (60 ms) for the compositor / app to
process it, then pastes the raw transcript. The preview hides
itself beforehand so the keystroke doesn't race focus
(existing Wayland-hardening path).
- settings.autoPaste = false → nothing was pasted in the first
place, so just overwrite the clipboard with raw. User's own
paste yields raw.
Backend: new paste_text_replacing Tauri command plus a mirror of the
paste-backend matrix for undo (wtype -M ctrl z / xdotool key ctrl+z /
ydotool keycodes 29:1 44:1 44:0 29:0 / osascript cmd+z / SendKeys '^z').
Reuses the pick_linux_backend_order Wayland-vs-X11 preference.
Registered in the Tauri command handler.
Acceptance per the brief: "after paste, Ctrl+Z within 5 s replaces
LLM output with raw transcript" — satisfied via the 4 s auto-hide
window on the preview's final phase. The click extends auto-hide so
the user actually sees the confirmation.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
prewarm_default_model loaded the model and returned. That moves the
model into RAM, but whisper.cpp still allocates its context window +
fills GPU shader caches on the first inference call — producing the
~4–5 s cold-start latency documented in ufal/whisper_streaming #96
and #135 that feels like "Kon dropped my first sentence."
Extend the pre-warm task: after engine.load, feed one second of
silence (16000 zero samples at 16 kHz) through transcribe_sync with
default options. Silence returns empty segments; the *work* is the
context allocation, which now happens at app boot rather than on the
user's first hotkey press.
Net: the user's first real dictation should complete within ~1.5× the
steady-state RTF they'll see on subsequent runs, satisfying the A.1
#23 acceptance criterion. No new public API; all inside the existing
spawn_blocking.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
