Commit graph

66 commits

Author SHA1 Message Date
Yun Chan
6fef05a39f fix(rag): don't carry a query, upload or indexing run across an account switch
The chunk store reads whatever database is open at call time, so a query that
was awaiting the model returned an answer built from the previous account's
documents, and a large file being read was inserted into the new account's
knowledge base and synced there. Each operation now remembers the account
scope it started in and stops if it changed.
2026-10-21 11:23:21 +09:00
Yun Chan
3e6cfd1ba8 fix(keybinding): release held shortcuts and modifiers when the screen locks or the PC sleeps
The lock desktop swallows keyup, so a hold-to-talk recording never ended,
the first shortcut after unlocking was dropped as auto-repeat, and Win from
Win+L made the next suggestion accept wait out the full 1.5s modifier timeout.
On powerMonitor lock-screen/suspend the chord machine now emits released for
emitted triggers and clears its state, and modifier tracking forgets held keys.
2026-10-21 00:28:40 +09:00
Yun Chan
8d003c8fca fix(caption): unload the caption-only STT model when a start is cancelled or fails
stop() during model loading saw no aux model yet and skipped the unload, and
the failure path (e.g. no microphone) never unloaded it, so the caption model
kept VRAM alongside the dictation model until the next session ended.
2026-10-21 00:25:35 +09:00
Yun Chan
c46713536d fix(rag): stop superseded indexing runs and refuse to wipe chunks on an empty reindex
- Each document gets an indexing run number; a reindex, synced replacement or
  delete makes the older run stop instead of embedding stale chunks, reporting
  'indexed' early and leaving a progress bar for a deleted document
- Reindexing a file that is now empty throws like addDocument instead of
  replacing the stored chunks with nothing
2026-10-21 00:24:38 +09:00
Yun Chan
45d8b1ddb1 fix(file-transcription): stop a Whisper repetition loop from spreading to every following chunk
Each 30-second chunk received the previous chunk's text as initial_prompt,
so once one chunk looped on a word the loop became the next chunk's hint
and the rest of the file was the same word. A looping chunk is now
re-transcribed without the hint, looped text is never carried forward, and
runs of 4+ identical phrases or segments are collapsed to one.
2026-10-20 22:56:40 +09:00
Yun Chan
95656b098a fix(screen-context): skip selected-text capture when the clipboard can't be fully restored
Capture still used the old snapshot code: a file copied in Explorer looked
empty and was cleared, and Excel's own formats were dropped on restore. It
now uses the shared clipboard snapshot policy and leaves the clipboard
alone when it couldn't put everything back.
2026-10-20 22:51:51 +09:00
Yun Chan
ce8d603b4a fix(chain): route chain steps through the LLM gateway instead of always calling the premium proxy
Chains were pinned to the premium service, so they always failed in
anonymous local mode and sent text to the cloud for signed-in users who had
chosen the local backend.
2026-10-20 22:48:17 +09:00
Yun Chan
3851c2a09a fix(conversation): drop the utterance when the user cancels while it is still being transcribed
Cancel only aborted an in-flight LLM response, which does not exist yet
during STT, so the cancelled speech was answered a few seconds later. A
turn counter bumped by cancel/stop now discards the late transcript.
2026-10-20 22:47:10 +09:00
Yun Chan
a03cda2211 fix(update): never offer a lower or equal version on prerelease channels
Setting a channel makes electron-updater allow downgrades, and the channel
check let prerelease channels accept them, so a stable install on beta would
be asked to 'update' down to an older beta.
2026-10-20 16:34:48 +09:00
Yun Chan
4e75f515e2 fix(dictionary): catch non-ASCII case duplicates when importing a dictionary file
SQLite lower() only folds ASCII, so importing 'äpfel' next to an existing
'Äpfel' created a second local entry that later collided with the server's
unique index and silently overwrote one spelling. Import now compares with
the same JS identity key as add/update, including duplicates within the file.
2026-10-20 16:34:12 +09:00
Yun Chan
c3ca95d20f fix(cloud-sync): finish active dictation and meeting recording before switching to another account's database
Signing in as a different account (e.g. while the previous one was still
waiting on an offline restore) swapped the database under a live
recording, so its final save and audio upload went into the new account.
The sign-out capture shutdown now also runs whenever the open database
belongs to another user.
2026-10-20 16:32:54 +09:00
Yun Chan
69c4307eee fix(sync): don't let a pull wipe the checkpointed transcript of a meeting this device is recording
Checkpoints stay local until stop, so a pull during recording replaced the
saved transcript with the server's empty one; a crash then recovered the
meeting as an error with nothing in it. Status, transcript, end time and
duration of a locally recording meeting are now kept on pull.
2026-10-20 16:31:08 +09:00
Yun Chan
fd9c97b7cc feat(release): one-command desktop release with Ed25519-signed update metadata
The installer is unsigned, so latest.yml is now signed with an Ed25519 key and the
app verifies the downloaded installer against it before installing.
npm run release -- patch|minor bumps, gates, tags, builds, signs, publishes,
re-verifies the public feed and pushes; --resume continues after a failure.
The retired GitLab registry publisher is removed.
2026-10-18 07:06:55 +09:00
Yun Chan
f5ae4c8c04 feat(desktop): propose typo fixes inside the suggestion popup and make suggestions snappier
While suggestions are on, typos are no longer replaced silently. The popup shows
the fix next to the candidate and applies it only when the candidate is accepted.
The per-minute request cap rises from 6 to 20 so suggestions keep appearing after
accepting one and continuing to type.
2026-10-18 07:06:55 +09:00
Yun Chan
d63854db67 fix(desktop): stop pinning local models in memory and cap audio buffers
Some checks failed
ci / 정본·보안·린트·타입·테스트 (push) Successful in 56s
ci / .NET API 서버 테스트 (push) Successful in 10s
deploy-site / deploy (push) Successful in 38s
ci / 모바일 린트·타입·Jest (push) Successful in 44s
ci / 워크스페이스 빌드 검증 (push) Successful in 54s
ci / Supabase Edge Functions + Cloudflare Worker (push) Successful in 22s
portable-unsigned / portable-windows (push) Failing after 35s
release / release-windows (push) Failing after 13s
The PC-wide slowdown was not a leak but model residency: boot warmup loaded
gemma4:e4b (~11.6 GB RAM incl. 5.9 GB pinned host buffer, no mmap on Windows
CUDA) and kept it for 30m, next to the large-v3 sidecar. Measured on the HEAD
build: free RAM 7.9 GB -> 32 MB within 20 s; after this change 10.8 GB with no
Ollama model loaded until typing starts.

- single LOCAL_LLM_KEEP_ALIVE (10m) for every local request; no boot warmup
- unload this app's resident Ollama models (keep_alive 0) on quit
- sidecar POST /unload for the aux slot; captions release it on stop
- partial dictation copies only the tail window, not the whole recording
- cap the voice-conversation listen buffer at 5 minutes
- mic test: generation guard so re-entry cannot orphan lease/timers
- startRealtime: single-flight and removeChannel to stop duplicate bindings
- meeting WAV save writes header and PCM without a concat copy

Docs: AI-01 policy note, GAP-PERF-01 with remaining work.
2026-10-11 11:35:15 +09:00
Yun Chan
3cac14f668 feat(desktop): typo auto-correct and context-aware sentence prediction
Some checks failed
ci / 정본·보안·린트·타입·테스트 (push) Has been cancelled
deploy-site / deploy (push) Has been cancelled
ci / 워크스페이스 빌드 검증 (push) Has been cancelled
ci / 모바일 린트·타입·Jest (push) Has been cancelled
ci / Supabase Edge Functions + Cloudflare Worker (push) Has been cancelled
ci / .NET API 서버 테스트 (push) Has been cancelled
Typo auto-correct (INPUT-19)
- TypoCorrectionService + core typo-correction: 2-set Korean layout
  mistypes both ways, always-wrong spellings, near-misses of the user's
  frequent words, optional model check re-validated in core; applied by
  backspace N + paste and cancelled on new keys or window change.

Prediction quality (INPUT-20/21)
- gemma models are sent raw with the user's unfinished sentence pre-filled
  into the model turn, so the model truly continues the text (mid-word
  joins and spacing exact). Other models keep the instruction prompt.
- Context: text before the caret (opening lines and a fact list of names,
  dates, amounts outside the 400-char window), messages just sent in the
  same chat window (memory only, 30 min), personal-graph follow-ups and
  IDF-ranked related sentences with Korean stemming.
- Guards: first sentence only, restated-prefix stripping (also fixes the
  instruction path), repeat drop, no invented dates/names/numbers
  (one retry), plain -다 register, no emoji unless the user uses them,
  search-box keywords. Fill loop mixes instruction prompts for variety.
- Opt-in real-model probe: tests/integration/suggestion-quality.ollama.test.ts

Suggestions now actually show up
- Terminal (Claude Code): caret trust from the input line, not the whole
  screen (status line below the box blocked every keystroke); Korean
  composition counts as typing at the end; a grey placeholder filling the
  empty box at once is not user text.
- Chrome: UIA focus flicker while typing no longer resets "edited".
- KakaoTalk/RichEdit: trailing paragraph mark ignored for edit position
  and stripped so accepted text keeps its separating space.
- While the overlay is visible the suggestion shortcuts are captured so
  Ctrl+Alt+Enter no longer also types a newline; released when hidden.
- nut-js preloaded after boot (first accept waited 4.8 s).
2026-10-03 21:48:28 +09:00
Yun Chan
5ca3709b32 feat: add mobile shell and unified search
Some checks failed
ci / 정본·보안·린트·타입·테스트 (push) Failing after 50s
ci / 워크스페이스 빌드 검증 (push) Has been skipped
ci / Supabase Edge Functions + Cloudflare Worker (push) Successful in 20s
ci / .NET API 서버 테스트 (push) Successful in 12s
deploy-site / deploy (push) Successful in 39s
ci / 모바일 린트·타입·Jest (push) Failing after 3h10m31s
2026-09-28 23:23:36 +09:00
Yun Chan
f505a03d34 fix(ci): point release-metadata checks at the relocated update and publish guards
Some checks are pending
ci / 정본·보안·린트·타입·테스트 (push) Waiting to run
ci / 워크스페이스 빌드 검증 (push) Blocked by required conditions
ci / 모바일 린트·타입·Jest (push) Waiting to run
ci / Supabase Edge Functions + Cloudflare Worker (push) Waiting to run
ci / .NET API 서버 테스트 (push) Waiting to run
deploy-site / deploy (push) Waiting to run
CI run 121 failed at "버전·계약 정본 대조": earlier refactors moved update
gating from UpdateService into update-policy.ts (evaluateUpdateOffer ties
decideUpdate + isWithinRollout), the differential-download switch into
update-adapters.ts, and the Forgejo re-release guard into
lib/immutable-package-guard.mjs (sha256 comparison, abort on conflict).
The checks still looked for the old markers in the old files. They now
verify the same guarantees where the code lives, and the self-test's
negative case targets assertNoImmutableConflicts.

Also clears the lint gate: a control-character regex in Keycap (range now
starts at U+0020) and unused helpers in two red-team tests.
2026-09-28 20:55:11 +09:00
Yun Chan
7b1210b5d1 feat(desktop): honest, unified dictation popups with Esc to cancel
Wave 3 of the 2026-09-28 design overhaul.

- Esc (no modifiers) cancels an in-progress dictation, recording or
  processing: VoiceModeService.requestEscape() queues the dead 'escape'
  action. Clicking the capsule cancels only while recording, so a stray
  click during processing no longer throws the text away.
- Errors show in the user's language: bootstrap now passes error.code to
  the tip (the translated branch never ran), and one core classifier
  (voice-error-message) feeds both the tip and the main-window alert.
- Capsule: bars stay flat without input (no fake "alive" wobble before the
  mic opens), an indeterminate bar replaces the made-up percentage, the
  partial transcript stays visible while processing, an Esc hint, and a
  brief "Inserted" state so success is visible.
- All six popups share popups/_shared (Pretendard, one panel radius,
  selection, kbd hints, reduced motion). History/command popups get
  titles, listbox roles and window sizes from popup-list-geometry (empty
  states were clipped); the command popup's "0" key is registered;
  result popup explains why it appeared and can be closed; captions keep
  a theme-independent dark backing on purpose.
- Popup themes follow the OS when set to system and the document lang
  follows the app language.
- Site hero demo mirrors the new capsule (done state, indeterminate bar,
  Esc hint; 10 locales).
2026-09-28 20:46:24 +09:00
Yun Chan
ba9ef9741e fix: red-team round 3 hardening across desktop, mobile, core and server
Batch of red-team r3 fixes that were in the working tree before the
2026-09-28 design overhaul, committed as one unit with their tests.

- desktop main: STT timeouts and sidecar, voice recording store, sync
  (credentials, audio, knowledge reindex, push gates), runtime
  provisioner, update policy, AltGr keybindings, voice-command policy,
  dictionary file codec/limits, meeting transcript condensing and a
  local recording ledger so interrupted-session recovery only closes
  meetings this device recorded (a phone's live meeting is left alone).
- mobile: login CSRF via implicit token callbacks rejected, account
  deletion/retention, durable queue retention, knowledge realtime
  without unfiltered DELETE, meeting re-record failure paths, cloud STT
  client, preferences store/resync.
- core: text chunking splits long unbroken transcripts to fit, template
  field policy, dictionary limits, meeting markdown inline handling.
- server: payple webhook policy and cancellation order scope, meeting
  document generation quota, team RPC null-role guard, unified LLM
  quota in-flight accounting, knowledge chunk vector index, meeting
  re-record failure paths (migrations 20260929*).
- ci: portable/runtime feed gates, update-policy schema, Forgejo file
  delete and alias planning.

Four older tests are updated to the new contracts rather than the old
behavior: token-pair auth callbacks are rejected, knowledge realtime no
longer subscribes to DELETE, long transcript lines are split, and
meeting recovery requires the local recording ledger for empty rows.
2026-09-28 20:45:52 +09:00
Yun Chan
1a4c39cb94 fix(keybinding): stop stealing chords on Windows, release mouse holds, keep cleared actions 2026-09-28 02:16:19 +09:00
Yun Chan
11bf995409 fix(runtime): release engine file holders right before swapping a reinstalled runtime 2026-09-28 02:16:19 +09:00
Yun Chan
88f3dbcd69 fix(update): re-evaluate policy on every update transition and skip NSIS updater on portable installs 2026-09-28 02:16:18 +09:00
Yun Chan
a4acb42573 fix(desktop): align runtime AppUserModelId with electron-builder appId 2026-09-28 02:16:18 +09:00
Yun Chan
5322d981cc fix(security): trust file: URLs only inside the app renderer directory 2026-09-28 02:16:18 +09:00
Yun Chan
2f94d24c99 fix(voice): stop mic-test ref stealing, unblock action queue, keep tail audio, and move context/LLM routing behind ports and pure policies 2026-09-28 02:16:16 +09:00
Yun Chan
2cd462333f fix(stt): gate local fallback on installed engine and extract SidecarSupervisor 2026-09-28 02:16:16 +09:00
Yun Chan
96b24e279c fix(text-insert): drop nut-js 300ms key delay and send paste as one chord 2026-09-28 02:16:15 +09:00
Yun Chan
9e5b94ced2 fix(suggestion): place overlay in DIPs, size it to content, and harden suggestion sessions 2026-09-28 02:16:15 +09:00
Yun Chan
75e053d72f fix(sync): keep desktop-only language and preset command across settings pulls 2026-09-28 02:16:15 +09:00
Yun Chan
ddc78546f0 fix(desktop): harden session, meeting, caption and LLM lifecycles; route LLM calls through the gateway 2026-09-28 02:16:15 +09:00
Yun Chan
3a46437f28 fix(sync,rag): keep restored rows, reconcile after tombstone pruning, reject partial knowledge docs, surface RAG indexing failures 2026-09-28 02:16:14 +09:00
Yun Chan
a85b24d385 fix(meeting): persist transcript segment edits, restore live recording view, and route settings deep links by tab id 2026-09-28 00:54:03 +09:00
Yun Chan
06e96c58f4 fix(sound): apply the Sound switch at once instead of after a restart 2026-09-28 00:53:47 +09:00
Yun Chan
91d4b4974c fix(tts): pass Windows TTS text via env and let stop() only cancel its own playback 2026-09-28 00:53:46 +09:00
Yun Chan
9d5d043e8b fix(license): keep signed offline keys across restarts and cloud sign-out 2026-09-28 00:53:45 +09:00
Yun Chan
524d390bc2 fix(runtime): stage runtime installs strictly, time out only on stalls, and make reinstall actually reinstall 2026-09-28 00:53:44 +09:00
Yun Chan
ba0dbbd813 refactor(instructions): share instruction template rendering via @d3ro/core 2026-09-28 00:53:44 +09:00
Yun Chan
d311e8123f fix(llm): keep system prompts on premium chat, reject incomplete streams, stop double-charging quota 2026-09-28 00:53:44 +09:00
Yun Chan
d96601a283 fix(file-transcription): provision ffmpeg before converting so clean installs stop failing with ENOENT 2026-09-28 00:53:43 +09:00
Yun Chan
c5b3bdedb6 fix(voice): let a click on the RecordingTip actually cancel the recording 2026-09-28 00:53:43 +09:00
Yun Chan
83cf9133ac refactor(desktop): session-scoped voice runtime, STT/LLM ports, caption ownership and meeting export fixes 2026-09-28 00:53:43 +09:00
Yun Chan
0adedf5e7b fix(text-insert): restore every clipboard format after paste and never wipe non-text clipboards 2026-09-28 00:53:42 +09:00
Yun Chan
190b6db284 refactor(suggestion): drive suggestion decisions from one core step and split SuggestionService collaborators 2026-09-28 00:53:42 +09:00
Yun Chan
9aa7302944 fix(sync): bind sync engine to its user DB, scope instructions/templates per account, harden navigation 2026-09-28 00:53:42 +09:00
Yun Chan
1b8fe445f3 fix(suggestion): stop generating mid-typing and let a pause bring a suggestion at once
With a 100 ms trigger left in the installed config, generation started in the
gaps between keystrokes; the next key ended the session silently, and that
session still counted against the 5 s minimum interval, so the moment the
user actually paused nothing came (rate-limited). The log showed a
generate-then-vanish cycle every 5-6 s.

The trigger delay now has a 500 ms floor and settings revision 6 resets a
stored value below it to 600 ms. A session ended because the user kept typing
no longer blocks the next request by the minimum interval (per-minute and
daily caps still apply), and that dismissal is logged.
2026-09-27 22:55:26 +09:00
Yun Chan
2aac10fc5d feat(desktop): send meeting transcript segments and preset prompt edits to the phone
The phone draws a meeting from its transcript segments before the edited
transcript, so desktop edits, auto-polish and diarization never showed there.
Every desktop transcript change now rebuilds the meeting's segments from its
[MM:SS] [speaker] lines and trims the rest; the line parser moves to
@d3ro/core/meeting-transcript and the meeting view uses it too.

Prompt edits of the four desktop presets that exist on the phone update the
server preset row (a reset restores its default; {{targetLanguage}} is sent
as English, the only target on both sides), and edits made on another desktop
come back. The free-prompt preset has no phone counterpart and stays local.
2026-09-27 16:24:25 +09:00
Yun Chan
9a8f7e6aa6 feat(desktop): sync knowledge, recordings and shared settings; play any recording
Knowledge documents travel as source-text chunks; each surface embeds them
with its own model, the server index is requested through embed-chunks, and
documents from the phone are stored without a file and indexed from their
chunks. Chunk text is now kept when local embedding fails, so reindexing no
longer needs the original file.

Recordings upload to the mobile storage contract (audio bucket under the
user's folder plus an audio_files row, 50 MiB cap, a Settings > Cloud
toggle) and are removed with their record. The history card gains a play
button that uses the local file or, for phone recordings, a signed URL.

Language (ko/en), system/light/dark theme, auto-polish and the active user
command follow the phone's user_settings with its revision rule; changes that
arrive from the phone reach the open window.
2026-09-27 14:44:56 +09:00
Yun Chan
0a4f5aee64 feat(desktop): two-way cloud sync with mobile and web
Rewrites the desktop mirror as services/sync/SyncEngine: a persistent
outbox, per-account server-clock keyset cursors with paging, pulls that never
overwrite unsent local edits, deletions both ways through sync_tombstones and
per-row failure isolation. It now covers history titles and favorites,
dictionary, every meeting's memos and documents, memo tags, user commands and
dictation/meeting templates, and registers the desktop as a device that the
phone can disconnect.

Fixes shipped defects: the first pull after sign-in fetched nothing, only
the first meeting's children were pushed, team meetings leaked into the
personal database and lost team_id on re-push, and Realtime never connected
because Electron's Node 20 has no global WebSocket (ws is now the transport).
Anonymous local-mode records are imported into the first account that signs
in. The settings sync section is translated and shows pending/rejected
changes; synced screens reload on app:dataChanged.
2026-09-27 14:04:49 +09:00
Yun Chan
0d92a4a853 fix(rag): do not mark a document indexed when no chunk could be embedded
Some checks failed
ci / 정본·보안·린트·타입·테스트 (push) Successful in 51s
ci / 모바일 린트·타입·Jest (push) Successful in 37s
ci / Supabase Edge Functions + Cloudflare Worker (push) Successful in 26s
ci / .NET API 서버 테스트 (push) Successful in 14s
deploy-site / deploy (push) Failing after 15s
ci / 워크스페이스 빌드 검증 (push) Failing after 11m7s
With the embedding server unavailable every chunk failed, yet the document
was stored as indexed=true with 0 chunks, so the knowledge base listed it as
searchable while queries could never match it. The red use-case test caught
this; it had been written off as an environment failure.

Now a run with zero embedded chunks leaves indexed=false and throws
RAGEmbeddingFailed (surfaced by reindex, logged by addDocument).

Tests that only hold on the Windows developer machine now declare it: the
bundled SoX binary and PowerShell device discovery run on win32 only, and
the sidecar venv test runs only when sidecar/.venv exists. The Linux Forgejo
runner skips them instead of failing.
2026-09-26 21:10:40 +09:00