docs: add v0.8.0 screenshots and update both READMEs (Paket L)

README.de.md

@@ -155,7 +155,7 @@ Blitztext registriert globale Hotkeys via `evdev`. Mit diesen Kombinationen hast
 Die KI-Workflows helfen bei Formulierung, Ton und Emojis. Die passenden Einstellungen findest du unter **Einstellungen → KI-Workflows**:
 
 <div align="center">
-  <img src="docs/screenshots/linux/en/settings-ai-workflows.png" alt="KI-Workflow-Einstellungen" width="480">
+  <img src="docs/screenshots/linux/settings-ai-workflows-en.png" alt="KI-Workflow-Einstellungen" width="480">
   <br><br>
 </div>
 
@@ -195,6 +195,12 @@ Für den Workflow **Blitztext+** (Text-Verbesserer) gibt es vorgefertigte Schrei
 
 Das **Compose-Fenster** (`✍ Compose…` im Tray-Kontextmenü) ermöglicht das Umschreiben beliebiger Texte mit der KI — ganz ohne Sprachaufnahme. Es eignet sich ideal zum Überarbeiten fertiger Entwürfe, E-Mails oder Notizen.
 
+<div align="center">
+  <br>
+  <img src="docs/screenshots/linux/compose-en.png" alt="Compose-Fenster" width="480">
+  <br><br>
+</div>
+
 **Öffnen:** Klick auf das Tray-Icon → **✍ Compose…**
 
 **Was du im Compose-Fenster tun kannst:**
@@ -251,7 +257,7 @@ Das Tray-Kontextmenü gibt dir schnellen Zugriff auf alle Workflows, das Compose
 
 <div align="center">
   <br>
-  <img src="docs/screenshots/linux/en/tray-menu.png" alt="Tray-Kontextmenü" width="280">
+  <img src="docs/screenshots/linux/tray-menu-en.png" alt="Tray-Kontextmenü" width="280">
   <br><br>
 </div>
 
@@ -266,11 +272,12 @@ Das Hauptfenster ist dein grafisches Kontrollzentrum — nützlich, wenn Hotkeys
 
 <div align="center">
   <br>
-  <img src="docs/screenshots/linux/en/blitztext-ready.png" alt="Hauptfenster" width="300">
+  <img src="docs/screenshots/linux/main-window-en.png" alt="Hauptfenster" width="300">
   <br><br>
 </div>
 
 - **Workflow-Dropdown:** Alle 5 Aufnahmemodi zur Auswahl.
+- **Schreibstil-Vorlage:** Sichtbar wenn **Blitztext+** gewählt ist — Preset direkt im Hauptfenster wählen. Änderungen werden sofort mit dem Tray synchronisiert.
 - **Start/Stopp-Button:** Klick zum Starten oder Beenden einer Aufnahme.
 - **Abbruch:** Bricht die aktuelle Aufnahme ohne Transkription ab.
 - **Diktat / Verlauf:** Schnellzugriff auf den Diktat-Modus und den Transkript-Verlauf.
@@ -286,8 +293,8 @@ Zusätzlich zu den Workflows bietet das Tool drei Komfort-Funktionen:
 
 <div align="center">
   <br>
-  <img src="docs/screenshots/linux/history.png" alt="Verlauf" width="340">
-  <img src="docs/screenshots/linux/en/read-aloud.png" alt="Vorlesen" width="340">
+  <img src="docs/screenshots/linux/history-en.png" alt="Verlauf" width="340">
+  <img src="docs/screenshots/linux/tts-en.png" alt="Vorlesen" width="340">
   <br><br>
 </div>
 
@@ -320,11 +327,11 @@ Alles wird lokal und sicher unter `~/.config/blitztext-linux/config.json` gespei
 Der Einstellungs-Dialog hat drei Tabs:
 
 <div align="center">
-  <img src="docs/screenshots/linux/en/settings-speech-recognition.png" alt="Einstellungen: Spracherkennung" width="480">
+  <img src="docs/screenshots/linux/settings-speech-en.png" alt="Einstellungen: Spracherkennung" width="480">
   <br><i>Spracherkennung — Whisper-Modell, Backend, Sprache, Hotkey-Modus und Aufnahmetaste.</i><br><br>
-  <img src="docs/screenshots/linux/en/settings-ai-workflows.png" alt="Einstellungen: KI-Workflows" width="480">
+  <img src="docs/screenshots/linux/settings-ai-workflows-en.png" alt="Einstellungen: KI-Workflows" width="480">
   <br><i>KI-Workflows — LLM-Anbieter, API-Key, Base-URL, Modell, Tonfall und Schreibstil-Vorlage.</i><br><br>
-  <img src="docs/screenshots/linux/en/settings-general.png" alt="Einstellungen: Allgemein" width="480">
+  <img src="docs/screenshots/linux/settings-general-en.png" alt="Einstellungen: Allgemein" width="480">
   <br><i>Allgemein — Auto-Paste, Diktat-Ordner, Verlaufsgröße, Sprache der Oberfläche und Signatur.</i><br><br>
 </div>
 

README.md

@@ -157,7 +157,7 @@ Blitztext registers global hotkeys via `evdev`. With these combinations you have
 The AI workflows help with phrasing, tone, and emojis. You'll find the relevant settings under **Settings → AI Workflows**:
 
 <div align="center">
-  <img src="docs/screenshots/linux/en/settings-ai-workflows.png" alt="AI workflow settings" width="480">
+  <img src="docs/screenshots/linux/settings-ai-workflows-en.png" alt="AI workflow settings" width="480">
   <br><br>
 </div>
 
@@ -197,6 +197,12 @@ For the **Blitztext+** workflow (text improver) there are ready-made writing-sty
 
 The **Compose window** (`✍ Compose…` in the tray menu) lets you rewrite any text using the AI — without recording your voice. It is ideal for editing existing drafts, emails, or notes.
 
+<div align="center">
+  <br>
+  <img src="docs/screenshots/linux/compose-en.png" alt="Compose window" width="480">
+  <br><br>
+</div>
+
 **How to open:** Click the tray icon → **✍ Compose…**
 
 **What you can do in the Compose window:**
@@ -253,7 +259,7 @@ The tray context menu gives you quick access to all workflows, the compose windo
 
 <div align="center">
   <br>
-  <img src="docs/screenshots/linux/en/tray-menu.png" alt="Tray context menu" width="280">
+  <img src="docs/screenshots/linux/tray-menu-en.png" alt="Tray context menu" width="280">
   <br><br>
 </div>
 
@@ -268,11 +274,12 @@ The main window is your graphical control center — useful when hotkeys are blo
 
 <div align="center">
   <br>
-  <img src="docs/screenshots/linux/en/blitztext-ready.png" alt="Main window" width="300">
+  <img src="docs/screenshots/linux/main-window-en.png" alt="Main window" width="300">
   <br><br>
 </div>
 
 - **Workflow dropdown:** Select from all 5 recording modes.
+- **Writing-style preset:** Visible when **Blitztext+** is selected — pick your preset directly in the main window. Changes sync to the tray instantly.
 - **Start/Stop button:** Click to begin or end a recording.
 - **Discard:** Cancels the current recording without transcription.
 - **Dictation / History:** Quick access to dictation mode and the transcript history.
@@ -288,8 +295,8 @@ In addition to the workflows, the tool offers three convenience functions:
 
 <div align="center">
   <br>
-  <img src="docs/screenshots/linux/history.png" alt="History" width="340">
-  <img src="docs/screenshots/linux/en/read-aloud.png" alt="Read aloud" width="340">
+  <img src="docs/screenshots/linux/history-en.png" alt="History" width="340">
+  <img src="docs/screenshots/linux/tts-en.png" alt="Read aloud" width="340">
   <br><br>
 </div>
 
@@ -323,11 +330,11 @@ Everything is stored locally and securely under `~/.config/blitztext-linux/confi
 The settings dialog has three tabs:
 
 <div align="center">
-  <img src="docs/screenshots/linux/en/settings-speech-recognition.png" alt="Settings: Speech Recognition" width="480">
+  <img src="docs/screenshots/linux/settings-speech-en.png" alt="Settings: Speech Recognition" width="480">
   <br><i>Speech Recognition — Whisper model, backend, language, hotkey mode, and recording key.</i><br><br>
-  <img src="docs/screenshots/linux/en/settings-ai-workflows.png" alt="Settings: AI Workflows" width="480">
+  <img src="docs/screenshots/linux/settings-ai-workflows-en.png" alt="Settings: AI Workflows" width="480">
   <br><i>AI Workflows — LLM provider, API key, base URL, model, tone, and writing-style preset.</i><br><br>
-  <img src="docs/screenshots/linux/en/settings-general.png" alt="Settings: General" width="480">
+  <img src="docs/screenshots/linux/settings-general-en.png" alt="Settings: General" width="480">
   <br><i>General — Auto-Paste, dictation folder, history size, interface language, and signature.</i><br><br>
 </div>
 

scripts/_make_screenshots.py

@@ -22,12 +22,15 @@
 from PyQt6.QtCore import Qt
 from PyQt6.QtWidgets import QApplication
 
+from app.blitztext_linux import BlitztextApp, Config, SettingsDialog
+from app.compose_window import ComposeWindow
 from app.config import BlitztextConfig
 from app.history_panel import HistoryPanel
 from app.i18n import set_language, t
+from app.llm_service import LLMService, WorkflowType
 from app.main_window import MainWindow
+from app.paste_service import PasteService
 from app.tts_window import TtsWindow
-from app.blitztext_linux import BlitztextApp, Config, SettingsDialog
 
 SCREENSHOT_DIR = Path("docs/screenshots/linux")
 CANVAS_SIZE = (1280, 640)
@@ -47,20 +50,22 @@
         "hero": "Your local AI voice assistant for KDE Plasma & Wayland",
         "sub": "Record speech, transcribe locally or online, optionally rewrite it with AI, and paste it directly into the active app.",
         "flow": "Record  •  Transcribe  •  Rewrite  •  Paste",
-        "feature_title": "What is new in v0.4.0",
+        "feature_title": "What is new in v0.8.0",
         "tag_new": "NEW",
         "chips": [
-            ("Multilingual UI", "Switch the whole app between English and German."),
-            ("Writing-style presets", "Ready-made Blitztext+ presets for common writing tasks."),
-            ("Offline-ready", "Local Whisper and privacy-friendly workflows stay available."),
-            ("Global hotkeys", "Capture dictation from anywhere in KDE Plasma."),
+            ("Compose window", "Draft, refine and compare AI-rewritten text before pasting it anywhere."),
+            ("Preset in main window", "Pick your writing style in the main window — changes sync to the tray instantly."),
+            ("Tone & template control", "Choose tone and writing template directly inside the Compose window."),
+            ("Prompt transparency", "Inspect and edit the AI system prompt before running a rewrite."),
         ],
         "labels": {
             "main": "Main window",
-            "general": "Settings → General",
+            "compose": "Compose window",
             "workflows": "Settings → AI Workflows",
             "tray": "Tray presets",
         },
+        "compose_input": "Please help me write a concise follow-up email for our product meeting earlier today.",
+        "compose_output": "Hi team,\n\nThank you for the productive discussion today. Here are the key action items we agreed on:\n\n• Finalise the API contract by Friday\n• Schedule a follow-up review for next Tuesday\n• Share the updated roadmap with stakeholders\n\nLet me know if I missed anything.\n\nBest,",
         "history_entries": [
             ("Please move tomorrow's team sync to 10:00.", False),
             ("Could you send me the updated rollout plan afterwards?", True),
@@ -74,20 +79,22 @@
         "hero": "Dein lokaler KI-Sprachassistent für KDE Plasma & Wayland",
         "sub": "Sprache aufnehmen, lokal oder online transkribieren, optional mit KI umformulieren und direkt in die aktive Anwendung einfügen.",
         "flow": "Aufnehmen  •  Transkribieren  •  Umformulieren  •  Direkt einfügen",
-        "feature_title": "Neu im Stand v0.4.0",
+        "feature_title": "Neu in v0.8.0",
         "tag_new": "NEU",
         "chips": [
-            ("Mehrsprachige Oberfläche", "Die komplette App lässt sich zwischen Deutsch und Englisch umschalten."),
-            ("Schreibstil-Vorlagen", "Vorgefertigte Blitztext+-Presets für typische Schreibaufgaben."),
-            ("Offline-fähig", "Lokale Whisper-Workflows bleiben für datensparsame Nutzung verfügbar."),
-            ("Globale Hotkeys", "Diktat direkt aus jeder Anwendung unter KDE Plasma starten."),
+            ("Entwurfsfenster", "Text entwerfen, verfeinern und KI-Varianten vergleichen, bevor du einfügst."),
+            ("Preset im Hauptfenster", "Schreibstil direkt im Hauptfenster wählen – Änderungen bleiben mit dem Tray synchron."),
+            ("Tonfall & Vorlage", "Tonfall und Schreibvorlage direkt im Entwurfsfenster steuern."),
+            ("Prompt-Transparenz", "KI-Systemprompt vor der Ausführung einsehen und anpassen."),
         ],
         "labels": {
             "main": "Hauptfenster",
-            "general": "Einstellungen → Allgemein",
+            "compose": "Entwurfsfenster",
             "workflows": "Einstellungen → KI-Workflows",
             "tray": "Tray-Presets",
         },
+        "compose_input": "Bitte hilf mir, eine knappe Nachfass-E-Mail zu unserem heutigen Produktmeeting zu schreiben.",
+        "compose_output": "Hallo zusammen,\n\nvielen Dank für die produktive Diskussion heute. Hier die vereinbarten Aufgaben:\n\n• API-Vertrag bis Freitag finalisieren\n• Review-Termin für nächsten Dienstag eintragen\n• Aktualisierte Roadmap an Stakeholder verteilen\n\nBitte meldet euch, falls ich etwas vergessen habe.\n\nViele Grüße,",
         "history_entries": [
             ("Bitte verschiebe das Team-Meeting morgen auf 10 Uhr.", False),
             ("Kannst du mir danach den aktualisierten Rollout-Plan schicken?", True),
@@ -277,14 +284,14 @@ def _make_banner(lang: str, out_dir: Path) -> None:
 
     screenshots = {
         "main": out_dir / f"main-window-{lang}.png",
-        "general": out_dir / f"settings-general-{lang}.png",
+        "compose": out_dir / f"compose-{lang}.png",
         "workflows": out_dir / f"settings-ai-workflows-{lang}.png",
         "tray": out_dir / f"tray-menu-{lang}.png",
     }
 
     placements = [
         (screenshots["main"], (824, 74), (220, 260), copy["labels"]["main"]),
-        (screenshots["general"], (1046, 74), (194, 260), copy["labels"]["general"]),
+        (screenshots["compose"], (1046, 74), (194, 260), copy["labels"]["compose"]),
         (screenshots["workflows"], (790, 350), (250, 236), copy["labels"]["workflows"]),
         (screenshots["tray"], (1054, 332), (186, 254), copy["labels"]["tray"]),
     ]
@@ -307,6 +314,36 @@ def _make_banner(lang: str, out_dir: Path) -> None:
         print(f"  ✓ {social_path.name}")
 
 
+class _FakeLLMService(LLMService):
+    """Minimal LLMService stub for offscreen rendering — never calls any API."""
+
+    def __init__(self) -> None:
+        self.api_key = "SCREENSHOT_DUMMY_TOKEN"
+        self.writing_preset = "standard"
+
+    def is_configured(self) -> bool:
+        return True
+
+    def rewrite_text(self, *args, **kwargs) -> str:  # type: ignore[override]
+        return ""
+
+    def build_system_prompt(self, *args, **kwargs) -> str:  # type: ignore[override]
+        return ""
+
+    def rewrite_raw(self, *args, **kwargs) -> str:  # type: ignore[override]
+        return ""
+
+
+class _FakePasteService(PasteService):
+    """Minimal PasteService stub — suppresses all clipboard / xdotool calls."""
+
+    def __init__(self) -> None:
+        pass
+
+    def paste(self, text: str, force_autopaste=None) -> None:
+        pass
+
+
 def _tab_index(tabs, key: str) -> int:
     """Resolve a settings tab by its i18n key, independent of tab order or language."""
     target = t(key)
@@ -336,9 +373,16 @@ def _render_language_set(out_dir: Path, lang: str) -> None:
             show_history_panel=lambda *a, **k: None,
             show_settings_dialog=lambda *a, **k: None,
             show_tts_window=lambda *a, **k: None,
+            main_window_preset_changed=lambda *a, **k: None,
         )
 
         main_window = MainWindow(controller)
+        # Switch to TEXT_IMPROVER so the writing-style preset combo is visible
+        for i in range(main_window._workflow_combo.count()):
+            if main_window._workflow_combo.itemData(i) == WorkflowType.TEXT_IMPROVER:
+                main_window._workflow_combo.setCurrentIndex(i)
+                _process_events()
+                break
         _grab(main_window, out_dir / f"main-window-{lang}.png")
         main_window.update_state("RECORDING", None, None)
         _grab(main_window, out_dir / f"main-window-recording-{lang}.png")
@@ -347,6 +391,8 @@ def _render_language_set(out_dir: Path, lang: str) -> None:
         settings = SettingsDialog(config)
         settings.tabs.setCurrentIndex(_tab_index(settings.tabs, "settings.tab.general"))
         _grab(settings, out_dir / f"settings-general-{lang}.png")
+        settings.tabs.setCurrentIndex(_tab_index(settings.tabs, "settings.tab.speech"))
+        _grab(settings, out_dir / f"settings-speech-{lang}.png")
         settings.tabs.setCurrentIndex(_tab_index(settings.tabs, "settings.tab.workflows"))
         _grab(settings, out_dir / f"settings-ai-workflows-{lang}.png")
         settings.close()
@@ -366,6 +412,12 @@ def _render_language_set(out_dir: Path, lang: str) -> None:
         _grab(tts, out_dir / f"tts-{lang}.png")
         tts.close()
 
+        compose = ComposeWindow(_FakeLLMService(), _FakePasteService(), config)
+        compose.set_input_text(copy["compose_input"])
+        compose.txtOutput.setPlainText(copy["compose_output"])
+        _grab(compose, out_dir / f"compose-{lang}.png")
+        compose.close()
+
         _capture_tray_menu(config, lang, out_dir / f"tray-menu-{lang}.png")
         _make_banner(lang, out_dir)