Calliop v0.1.8

Calliop v0.1.8

Feature release — configurable cleanup levels (Off / Light / Full), pause punctuation presets, and a polished Achievements screen.

Changes

Dictation quality

• Cleanup levels in Settings — choose Verbatim (off), Light (deterministic fillers + pause punctuation, no LLM), or Full (local AI cleanup)
• Pause punctuation presets — Fast, Natural, or Formal to match your speaking pace and desired text structure
• Existing auto_edit preference migrates automatically to the new mode system

Achievements UI

• Redesigned Achievements screen with hero band, category tabs, status filters, and loading skeletons
• Shared AchievementsContext for unlock notifications and view state
• Refined unlock queue animations and tier styling

Benchmarks

• Expanded French STT benchmark corpus and improved benchmark-stt reporting

Installation (Windows)

1. Download Calliop_*_x64-setup.exe (NSIS) or .msi.
2. Install WebView2 if needed.
3. Accept the SmartScreen warning (unsigned binary in v1).
4. On first launch, the Whisper model downloads (~466 MB).

See docs/installation.md.

macOS / Linux (experimental)

DMG, AppImage, and deb artifacts are provided unsigned and untested for v1. Use at your own risk.

Known limitations

• Windows binaries are unsigned (SmartScreen)
• ML models download separately on first use
• Automatic updates are enabled by default; disable in Settings → Advanced if needed
• No official macOS/Linux support in v1

License

AGPL-3.0

Calliop v0.1.7

Calliop v0.1.7

Feature release — achievements and gamification, a more reliable dictation pipeline, and improved LLM cleanup fidelity.

Changes

Achievements

• Achievements view in the sidebar with category filters, progress bar, and unlock cards (EN/FR)
• Unlock queue — toast-style notifications when new achievements are earned
• Local tracking — milestones, streaks, speed, explorer, learner, and secret achievements evaluated offline from dictation activity
• Achievements unlock as you dictate, explore settings, and build habits — no cloud account required

Dictation & pipeline

• Whisper engine readiness — clearer handling when the STT model is loading or unavailable, with PipelineBusy feedback instead of silent failures
• Deferred dictation — hotkey presses while the pipeline is busy are queued and retried when ready
• Pause punctuation — softer handling of mid-dictation pauses so transcripts read more naturally
• Hardened pipeline path from hotkey through text injection, with improved error recovery

LLM cleanup

• Fidelity-first system prompt — cleanup preserves your wording more closely while still fixing fillers and punctuation
• More reliable fallback when the model returns empty or invalid output

UI

• Restored toolbar actions, stagger reveal animation, and tooltip cleanup on the main view

Installation (Windows)

1. Download Calliop_*_x64-setup.exe (NSIS) or .msi.
2. Install WebView2 if needed.
3. Accept the SmartScreen warning (unsigned binary in v1).
4. On first launch, the Whisper model downloads (~466 MB).

See docs/installation.md.

macOS / Linux (experimental)

DMG, AppImage, and deb artifacts are provided unsigned and untested for v1. Use at your own risk.

Known limitations

• Windows binaries are unsigned (SmartScreen)
• ML models download separately on first use
• Automatic updates are enabled by default; disable in Settings → Advanced if needed
• No official macOS/Linux support in v1

License

AGPL-3.0

Calliop v0.1.6

Calliop v0.1.6

Feature release — redesigned Settings modal, richer Insights dashboard, new Whisper models, autostart on login, and more reliable LLM cleanup.

Changes

Settings

• Settings modal with sidebar navigation, section routing, and accessible labels (EN/FR)
• Updates panel in Settings to check for and download app updates from the UI
• Autostart on login — enabled by default; toggle in Settings
• Auto-edit enabled by default for new installs, with a one-time migration for existing users

Insights

• Activity streaks and estimated time saved metrics
• Hour × app heatmap and animated metric cards for a clearer usage overview
• Responsive layout polish and loading skeletons across Insight tabs

Models

• Distil FR v0.2 Whisper options: dec16 Q5_0 and dec16 Q8_0
• Updated model labels for Qwen3.5 across locales and the model catalog

Dictation & cleanup

• LLM cleanup fallback when the model returns empty output, thinking-only blocks, or empty quotes
• Stronger output validation with clearer error states for worker and validation failures

Installation (Windows)

1. Download Calliop_*_x64-setup.exe (NSIS) or .msi.
2. Install WebView2 if needed.
3. Accept the SmartScreen warning (unsigned binary in v1).
4. On first launch, the Whisper model downloads (~466 MB).

See docs/installation.md.

macOS / Linux (experimental)

DMG, AppImage, and deb artifacts are provided unsigned and untested for v1. Use at your own risk.

Known limitations

• Windows binaries are unsigned (SmartScreen)
• ML models download separately on first use
• Automatic updates are enabled by default; disable in Settings → Advanced if needed
• No official macOS/Linux support in v1

License

AGPL-3.0

Calliop v0.1.5

Calliop v0.1.5

Patch release — STT segment progress during long dictation, Qwen3.5 4B model migration, clearer installed-models panel, and more reliable hotkeys while models load.

Changes

Dictation progress

• STT segment progress — live count of transcribed segments in the main view and pipeline status card during long recordings (EN/FR)

Settings & models

• Installed models section in Settings lists downloaded Whisper and LLM weights with streamlined actions
• Qwen3.5 4B replaces Qwen3 4B as the large LLM option (~2.7 GB)
• Automatic migration removes legacy Qwen3 4B files after upgrade
• Stronger model validation and updated labels across EN/FR locales

Hotkeys

• HotkeySuspendGuard suspends shortcut registration while models load, then restores the dictation hotkey from settings
• Improved hotkey registration error handling during startup

Installation (Windows)

1. Download Calliop_*_x64-setup.exe (NSIS) or .msi.
2. Install WebView2 if needed.
3. Accept the SmartScreen warning (unsigned binary in v1).
4. On first launch, the Whisper model downloads (~466 MB).

See docs/installation.md.

macOS / Linux (experimental)

DMG, AppImage, and deb artifacts are provided unsigned and untested for v1. Use at your own risk.

Known limitations

• Windows binaries are unsigned (SmartScreen)
• ML models download separately on first use
• Automatic updates are enabled by default; disable in Settings → Advanced if needed
• No official macOS/Linux support in v1

License

AGPL-3.0

Calliop v0.1.4

Calliop v0.1.4

Feature release — live editing during dictation, update notifications, richer dictation controls, and model management in Settings.

Changes

Live dictation

• Live editing pipeline — punctuation on pause, list formatting, and pipelined LLM cleanup while you dictate
• Improved list formatting in transcript processing
• Stronger LLM cleanup validation so prompt fixes apply without rebuilding the sidecar

Dictation controls

• Clearer toggle vs push-to-talk behavior with busy hints during transcription and injection
• Cancel during transcription by releasing the shortcut (no text injected)
• Updated user guide with a shortcut action table (EN)

Updates

• Update notifications and background download with in-app toasts
• Streamlined update check logic

Settings & models

• Delete and reinstall models from Settings (Whisper and LLM)
• Better handling and messaging for corrupted model files (EN/FR)
• New model management actions and status feedback in the Models panel

Dictionary & injection

• Dictionary learning ignores appended tokens and case-only changes to reduce pollution
• More reliable paste injection error handling on Windows

Hotkeys & LLM tooling

• Hotkey capture includes modifier keys for more accurate shortcut mapping
• test-llm supports raw input and model selection for easier debugging

Installation (Windows)

1. Download Calliop_*_x64-setup.exe (NSIS) or .msi.
2. Install WebView2 if needed.
3. Accept the SmartScreen warning (unsigned binary in v1).
4. On first launch, the Whisper model downloads (~466 MB).

See docs/installation.md.

macOS / Linux (experimental)

DMG, AppImage, and deb artifacts are provided unsigned and untested for v1. Use at your own risk.

Known limitations

• Windows binaries are unsigned (SmartScreen)
• ML models download separately on first use
• Automatic updates are enabled by default; disable in Settings → Advanced if needed
• No official macOS/Linux support in v1

License

AGPL-3.0

Calliop v0.1.3

Calliop v0.1.3

Patch release — improved hotkey capture, settings polish, copy feedback, and an audio waveform fix.

Changes

Hotkeys & settings

• Support modifier-only hotkeys and the Windows key in shortcut capture
• Reject unsupported keys with clearer error messages (EN/FR)
• Capture hotkeys on key-up for a more reliable recording experience
• Show the application version in Settings (English and French)
• Display low-power mode hints when downloading models
• Notify on model download failures for both Whisper and LLM sidecar
• Streamline Settings by removing the redundant progress bar

History & copy

• New CopyButton with animated feedback in history and snippet variables
• Clearer per-entry action state when copying or reinjecting from history

UI

• Smoother ProgressBar animation with reduced-motion support

Audio

• Fix band-level calculation in the overlay waveform pipeline

Installation (Windows)

1. Download Calliop_*_x64-setup.exe (NSIS) or .msi.
2. Install WebView2 if needed.
3. Accept the SmartScreen warning (unsigned binary in v1).
4. On first launch, the Whisper model downloads (~466 MB).

See docs/installation.md.

macOS / Linux (experimental)

DMG, AppImage, and deb artifacts are provided unsigned and untested for v1. Use at your own risk.

Known limitations

• Windows binaries are unsigned (SmartScreen)
• ML models download separately on first use
• Automatic updates are enabled by default; disable in Settings → Advanced if needed
• No official macOS/Linux support in v1

License

AGPL-3.0

Calliop v0.1.2

Calliop v0.1.2

Patch release — fixes auto-update signing and enables updates by default.

Changes

• Fix updater public key alignment between tauri.conf.json and CI signing key
• Enable automatic updates by default (still disableable in Settings → Advanced)
• Restart the app automatically after a successful signed update
• Show a desktop notification when an update download or install fails
• CI release workflow verifies updater pubkey before building

Important for v0.1.1 users

Installations from v0.1.1 were built with a mismatched updater public key and cannot auto-update. Please reinstall manually once from this release; later updates will then work automatically.

Installation (Windows)

1. Download Calliop_*_x64-setup.exe (NSIS) or .msi.
2. Install WebView2 if needed.
3. Accept the SmartScreen warning (unsigned binary in v1).
4. On first launch, the Whisper model downloads (~466 MB).

See docs/installation.md.

macOS / Linux (experimental)

DMG, AppImage, and deb artifacts are provided unsigned and untested for v1. Use at your own risk.

Known limitations

• Windows binaries are unsigned (SmartScreen)
• ML models download separately on first use
• Automatic updates can be disabled in Settings → Advanced
• No official macOS/Linux support in v1

License

AGPL-3.0

Calliop v0.1.1

Calliop v0.1.1

Patch release — CI/release fixes and minor improvements since v0.1.0.

Changes

• Fix GitHub release workflow (signing key password, release notes loading)
• README and documentation cleanup
• Tray menu localization fix (ampersand handling on Windows)
• CI rustfmt fix

Installation (Windows)

1. Download Calliop_*_x64-setup.exe (NSIS) or .msi.
2. Install WebView2 if needed.
3. Accept the SmartScreen warning (unsigned binary in v1).
4. On first launch, the Whisper model downloads (~466 MB).

See docs/installation.md.

macOS / Linux (experimental)

DMG, AppImage, and deb artifacts are provided unsigned and untested for v1. Use at your own risk.

Known limitations

• Windows binaries are unsigned (SmartScreen)
• ML models download separately on first use
• Automatic updates: opt-in in Settings → Advanced
• No official macOS/Linux support in v1

License

AGPL-3.0