docs(wiki): document the floating cheatsheet + clean Rules-and-Dictionary diagram - Voice-Commands (EN+FR): nouvelle section "Floating cheatsheet" / "Aide-mémoire flottant" avec capture, 4 façons de l'ouvrir (raccourci, plasmoid, tray, CLI), table des contrôles fenêtre (drag, redim 4 bords, zoom texte sans changer la fenêtre, opacité). - Plasmoid-Widget (EN+FR): bouton 🗒 Cheatsheet ajouté au footer du popup. - Tray-Icon (EN+FR): item "Toggle voice commands cheatsheet" ajouté au footer du menu contextuel. - Keyboard-Shortcuts (EN+FR): nouvelle ligne dans la table des raccourcis par défaut (Ctrl+Alt+F9 = aide-mémoire) + paragraphe qui explique le combo "Same key + Alt/Ctrl/Ctrl+Alt/Shift" (mécanisme unifié avec le raccourci traduction). - Rules-and-Dictionary (EN+FR): retrait du diagramme Mermaid "Spoken → Regex match? → Output" qui n'apportait rien et alourdissait la page sans clarifier le propos.

rcspam committed Apr 26, 2026

docs(wiki): bump window screenshots from 720 to 960 across all pages

rcspam committed Apr 24, 2026

docs(wiki): tray clicks — left toggles start/stop, middle-click or ESC cancels

rcspam committed Apr 24, 2026

docs(wiki): tray is enabled via dictee-setup Visual feedback tab (add screenshot)

rcspam committed Apr 24, 2026

docs(wiki): rewrite Tray-Icon as user-focused (drop dev internals) Match the Plasmoid-Widget treatment: - drop theme detection algorithm (Python snippet), cross-process sync internals, SVG filenames in status table, "fix landed in v1.3-beta4" commit references, detailed xdg-desktop-portal / waybar JSON snippets. - restructure around user actions: clicks (left / ctrl+left / middle / right), then the actual right-click menu verbatim from dictee-tray.py (actions → backends → files & meetings → toggles → footer). - use Keep meeting active and real toggle labels (LLM / Audio context / Short text fix), matching what the user actually sees. - replace generic troubleshooting with action-oriented bullets. - bump hero screenshot to 720px like other pages.

rcspam committed Apr 24, 2026

docs(wiki): rename FR pages with fr- prefix for auto-sidebar grouping Rename all 20 French pages from <Name>-fr.md to fr-<Name>.md so they cluster together in GitHub Wiki's alphabetical Pages list instead of being interspersed with English pages by topic. Updated all internal links across 41 files: - Cross-language switchers at top of pages - 'Étapes suivantes' / 'Next steps' link lists - Inline links in page bodies - _Sidebar.md FR link Python script replaces '<Name>-fr' with 'fr-<Name>' in order of decreasing length to avoid partial matches. No manual edits.

rcspam committed Apr 21, 2026

docs(wiki): inline SVG icons directly in Tray-Icon status table (EN+FR)

rcspam committed Apr 21, 2026

docs(wiki): fix Canary 25-lang + real tray icons (not emojis) Canary-1B v2 supports 25 European languages with translation ↔ English (48 pairs), not 4 languages as previously stated. The 4-language claim was from Canary v1 — dictee ships v2 since v1.2.0. Pages updated (EN + FR): - ASR-Backends: table + decision matrix + Canary section trade-offs - Translation: comparison table + Canary-1B built-in section + pairs table + target-language count mention - Changelog: v1.2.0 feature description Tray-Icon pages: replace emoji placeholders with real parakeet-*.svg icons from /usr/share/dictee/icons/: - parakeet-active(.svg/.dark.svg) = idle/ready - parakeet-recording.svg = recording - parakeet-transcribing.svg = transcribing - parakeet-diarize.svg = meeting mode - parakeet-inactive(.svg/-dark.svg) = stopped - parakeet-offline.svg = unavailable Each icon served via raw.githubusercontent.com for preview in wiki.

rcspam committed Apr 21, 2026

docs(wiki): fill wave V4-fr — Plasmoid, Tray, Keyboard-Shortcuts, Diarization

rcspam committed Apr 21, 2026

docs(wiki): fill wave V5 + add hero screenshots on V1-V4 Reference pages fully written (~1500 lines total): CLI-Reference (~300 lines) - Every binary: shell, Python, Rust - Full flag tables - Environment variables (dictation/PP/backends/runtime) - Exit codes - State files with UID suffix Troubleshooting (~320 lines) - By category: install / GPU / audio / daemon / PP / UI - Symptom + cause + fix for each - Log collection recipe - Per-component journalctl paths FAQ (~270 lines) - Design: why Rust, why not streaming, why ONNX, why Parakeet - Features: offline, multi-backend, Wayland, Flatpak - Security: multi-user, LLM privacy, model telemetry - Comparisons: nerd-dictation, Talon, Whisper.cpp - Hardware: minimum, AMD/Intel, Apple Silicon Developer-Guide (~340 lines) - Full repo layout - Cargo features matrix - Build scripts (.deb/.rpm/PKGBUILD) - Audio pipeline internals - Socket protocol - 682 + 148 test counts - gettext + plasmoid translations workflow - Contributing: branching, commits, adding ASR/lang Changelog (~250 lines) - v1.3.0 RC1/beta5/beta1-4 detailed - v1.2.0 (backend unification, onnx-asr removal) - v1.1.0 (translation backends) - v1.0.0 (initial) - Roadmap v1.3 final (auto-chunking, tests, restart) - Roadmap v1.4+ (Whisper translate, hotwords, Moonshine, CLI STT, VAD, streaming, overlay, Flatpak, GNOME/COSMIC) Hero screenshots added to V1-V4 pages via GitHub raw URLs: - Installation, ASR-Backends, Translation → setup screens - Post-Processing, Rules-and-Dictionary → regex editor + test panel - Plasmoid, Tray, Keyboard-Shortcuts, Diarization → VM screenshots

rcspam committed Apr 21, 2026

docs(wiki): fill wave V4 — Plasmoid, Tray, Keyboard-Shortcuts, Diarization UX + speaker pages fully written (total ~1100 lines): Plasmoid-Widget - Install from package / tarball - Compact vs full representation - 5 animation styles with tuning options - Language selector (dynamic per-backend) - LLM/Short/Meeting cross-process toggles (v1.3) - Theme detection (v1.3-rc1 Fedora fix) - Known quirks (popup orphan, elide fix) Tray-Icon - When to use tray vs plasmoid (decision matrix) - SNI vs AppIndicator3 fallback - Systemd user service autostart - Full menu structure (ASR/trans backends, toggles, daemon) - Status icons per state - Tiling WM integration (waybar SNI) - GNOME extension requirement Keyboard-Shortcuts - Default F9 (not Super+Space) - 3 capture mechanisms (KDE kglobalaccel, GNOME gsettings, dictee-ptt) - Setup Wizard capture handshake (pause dictee-ptt for Qt grab) - Per-WM config snippets (Sway, i3, Hyprland) - dictee-ptt toggle/hold/double-tap modes - Input group requirement - Keycode reference (F1-F12, ESC) Diarization - Sortformer architecture - 4-speaker cap + quality per speaker count - VRAM matrix: streaming / combined / chunked (v1.3 final) - Batch chunked prototype validation (54min in 122s) - CLI binaries comparison table - Output formats (plain/rttm/json/srt) - Why PP is bypassed (per-step rationale) - Limitations (anonymous labels, silent CPU fallback)

rcspam committed Apr 21, 2026

docs(wiki): scaffold 20 pages (stubs) Initial wiki structure for dictee v1.3.0 documentation: Getting started - Home (landing + TOC + 3 entry paths) - Installation (1-liner, .deb/.rpm/AUR/tarball, aarch64) - GPU-Setup (CUDA prereqs per distro + fallback) Speech recognition - ASR-Backends (comparison 4 backends) - Parakeet-TDT-Deep-Dive (main model, 25 langs, VRAM limits) Translation - Translation (5 backends + privacy matrix) - Ollama-Setup (install + Gemma 3 4B + prompts) User interfaces - Plasmoid-Widget (KDE Plasma 6 + 5 animations) - Tray-Icon (PyQt6 + SNI + AppIndicator fallback) - Keyboard-Shortcuts (KDE/GNOME/tiling WMs + double tap) Post-processing pipeline - Post-Processing-Overview (12 steps + diagram) - Rules-and-Dictionary (regex + dict + 7 langs) - LLM-Correction (first/last/hybrid positions) - Numbers-Dates-Continuation (cardinal/ordinal/version/buffer) Diarization & CLI - Diarization (Sortformer + 4 speakers + v1.4 batch) - CLI-Reference (every command + env vars + exit codes) Reference - Troubleshooting (errors by category + log locations) - FAQ (design + features + comparisons) - Developer-Guide (build + tests + i18n + contrib) - Changelog (version history) All 20 pages are stubs: H1 title + intro + TOC + minimal reference table. Content will be filled in 5 waves (V1-V5) per docs/plan-readme-wiki-v1.3.md.

rcspam committed Apr 21, 2026