pygpt-net 2.6.30__py3-none-any.whl → 2.6.31__py3-none-any.whl

pygpt_net/data/config/models.json

@@ -1,8 +1,8 @@
 {
     "__meta__": {
-        "version": "2.6.30",
-        "app.version": "2.6.30",
-        "updated_at": "2025-08-29T23:07:35"
+        "version": "2.6.31",
+        "app.version": "2.6.31",
+        "updated_at": "2025-09-01T23:07:35"
     },
     "items": {
         "SpeakLeash/bielik-11b-v2.3-instruct:Q4_K_M": {
@@ -873,6 +873,44 @@
             "provider": "google",
             "tool_calls": true
         },
+        "gemini-2.5-flash-preview-native-audio-dialog": {
+            "id": "gemini-2.5-flash-preview-native-audio-dialog",
+            "name": "gemini-2.5-flash-preview-native-audio-dialog",
+            "mode": [
+                "audio"
+            ],
+            "llama_index": {
+                "args": [
+                    {
+                        "name": "model",
+                        "value": "models/gemini-2.5-flash-preview-native-audio-dialog",
+                        "type": "str"
+                    }
+                ],
+                "env": [
+                    {
+                        "name": "GOOGLE_API_KEY",
+                        "value": "{api_key_google}",
+                        "type": "str"
+                    }
+                ]
+            },
+            "ctx": 128000,
+            "tokens": 8000,
+            "default": false,
+            "input": [
+                "text",
+                "audio"
+            ],
+            "output": [
+                "text",
+                "audio"
+            ],
+            "extra": {},
+            "imported": true,
+            "provider": "google",
+            "tool_calls": true
+        },
         "gemini-2.5-pro": {
             "id": "gemini-2.5-pro",
             "name": "gemini-2.5-pro",
@@ -1457,55 +1495,6 @@
             "provider": "openai",
             "tool_calls": true
         },
-        "gpt-4o-audio-preview": {
-            "id": "gpt-4o-audio-preview",
-            "name": "gpt-4o-audio-preview",
-            "mode": [
-                "audio"
-            ],
-            "llama_index": {
-                "args": [
-                    {
-                        "name": "model",
-                        "value": "gpt-4o-audio-preview",
-                        "type": "str"
-                    }
-                ],
-                "env": [
-                    {
-                        "name": "OPENAI_API_KEY",
-                        "value": "{api_key}"
-                    },
-                    {
-                        "name": "OPENAI_API_BASE",
-                        "value": "{api_endpoint}"
-                    },
-                    {
-                        "name": "AZURE_OPENAI_ENDPOINT",
-                        "value": "{api_azure_endpoint}"
-                    },
-                    {
-                        "name": "OPENAI_API_VERSION",
-                        "value": "{api_azure_version}"
-                    }
-                ]
-            },
-            "ctx": 128000,
-            "tokens": 16384,
-            "default": false,
-            "input": [
-                "text",
-                "audio"
-            ],
-            "output": [
-                "text",
-                "audio"
-            ],
-            "extra": {},
-            "imported": false,
-            "provider": "openai",
-            "tool_calls": true
-        },
         "gpt-4o-mini": {
             "id": "gpt-4o-mini",
             "name": "gpt-4o-mini",
@@ -1561,55 +1550,6 @@
             "provider": "openai",
             "tool_calls": true
         },
-        "gpt-4o-mini-audio-preview": {
-            "id": "gpt-4o-mini-audio-preview",
-            "name": "gpt-4o-mini-audio-preview",
-            "mode": [
-                "audio"
-            ],
-            "llama_index": {
-                "args": [
-                    {
-                        "name": "model",
-                        "value": "gpt-4o-mini-audio-preview",
-                        "type": "str"
-                    }
-                ],
-                "env": [
-                    {
-                        "name": "OPENAI_API_KEY",
-                        "value": "{api_key}"
-                    },
-                    {
-                        "name": "OPENAI_API_BASE",
-                        "value": "{api_endpoint}"
-                    },
-                    {
-                        "name": "AZURE_OPENAI_ENDPOINT",
-                        "value": "{api_azure_endpoint}"
-                    },
-                    {
-                        "name": "OPENAI_API_VERSION",
-                        "value": "{api_azure_version}"
-                    }
-                ]
-            },
-            "ctx": 128000,
-            "tokens": 16384,
-            "default": false,
-            "input": [
-                "text",
-                "audio"
-            ],
-            "output": [
-                "text",
-                "audio"
-            ],
-            "extra": {},
-            "imported": false,
-            "provider": "openai",
-            "tool_calls": true
-        },
         "gpt-5": {
             "id": "gpt-5",
             "name": "gpt-5 (medium)",
@@ -2303,6 +2243,112 @@
             "provider": "ollama",
             "tool_calls": true
         },
+        "gpt-realtime": {
+            "id": "gpt-realtime",
+            "name": "gpt-realtime",
+            "mode": [
+                "audio"
+            ],
+            "llama_index": {
+                "args": [
+                    {
+                        "name": "model",
+                        "value": "gpt-realtime",
+                        "type": "str"
+                    }
+                ],
+                "env": [
+                    {
+                        "name": "OPENAI_API_KEY",
+                        "value": "{api_key}",
+                        "type": "str"
+                    },
+                    {
+                        "name": "OPENAI_API_BASE",
+                        "value": "{api_endpoint}",
+                        "type": "str"
+                    },
+                    {
+                        "name": "AZURE_OPENAI_ENDPOINT",
+                        "value": "{api_azure_endpoint}",
+                        "type": "str"
+                    },
+                    {
+                        "name": "OPENAI_API_VERSION",
+                        "value": "{api_azure_version}",
+                        "type": "str"
+                    }
+                ]
+            },
+            "ctx": 32000,
+            "tokens": 4096,
+            "default": true,
+            "input": [
+                "text",
+                "audio"
+            ],
+            "output": [
+                "text",
+                "audio"
+            ],
+            "extra": {},
+            "imported": false,
+            "provider": "openai",
+            "tool_calls": true
+        },
+        "gpt-4o-realtime-preview": {
+            "id": "gpt-4o-realtime-preview",
+            "name": "gpt-4o-realtime-preview",
+            "mode": [
+                "audio"
+            ],
+            "llama_index": {
+                "args": [
+                    {
+                        "name": "model",
+                        "value": "gpt-4o-realtime-preview",
+                        "type": "str"
+                    }
+                ],
+                "env": [
+                    {
+                        "name": "OPENAI_API_KEY",
+                        "value": "{api_key}",
+                        "type": "str"
+                    },
+                    {
+                        "name": "OPENAI_API_BASE",
+                        "value": "{api_endpoint}",
+                        "type": "str"
+                    },
+                    {
+                        "name": "AZURE_OPENAI_ENDPOINT",
+                        "value": "{api_azure_endpoint}",
+                        "type": "str"
+                    },
+                    {
+                        "name": "OPENAI_API_VERSION",
+                        "value": "{api_azure_version}",
+                        "type": "str"
+                    }
+                ]
+            },
+            "ctx": 32000,
+            "tokens": 4096,
+            "default": false,
+            "input": [
+                "text",
+                "audio"
+            ],
+            "output": [
+                "text",
+                "audio"
+            ],
+            "extra": {},
+            "imported": true,
+            "provider": "openai",
+            "tool_calls": true
+        },
         "grok-2-vision": {
             "id": "grok-2-vision",
             "name": "grok-2-vision",
@@ -2597,7 +2643,7 @@
                 "image"
             ],
             "extra": {},
-            "imported": true,
+            "imported": false,
             "provider": "google",
             "tool_calls": true
         },
@@ -2633,7 +2679,7 @@
                 "image"
             ],
             "extra": {},
-            "imported": true,
+            "imported": false,
             "provider": "google",
             "tool_calls": true
         },

pygpt_net/data/config/settings.json

@@ -1501,6 +1501,30 @@
         "advanced": false,        
         "tab": "options"
     },    
+    "audio.input.vad.prefix": {
+        "section": "audio",
+        "type": "int",
+        "slider": false,
+        "label": "settings.audio.input.vad.prefix",
+        "value": 300,
+        "min": 0,
+        "multiplier": 1,
+        "step": 1,
+        "advanced": false,
+        "tab": "options"
+    },
+    "audio.input.vad.silence": {
+        "section": "audio",
+        "type": "int",
+        "slider": false,
+        "label": "settings.audio.input.vad.silence",
+        "value": 2000,
+        "min": 0,
+        "multiplier": 1,
+        "step": 1,
+        "advanced": false,
+        "tab": "options"
+    },
     "audio.cache.enabled": {
         "section": "audio",
         "type": "bool",
@@ -1655,6 +1679,20 @@
         "advanced": false,
         "tab": "Google"
     },
+    "remote_tools.google.url_ctx": {
+        "section": "remote_tools",
+        "type": "bool",
+        "slider": false,
+        "label": "settings.remote_tools.google.url_ctx",
+        "description": "settings.remote_tools.google.url_ctx.desc",
+        "value": true,
+        "min": null,
+        "max": null,
+        "multiplier": null,
+        "step": null,
+        "advanced": false,
+        "tab": "Google"
+    },
     "llama.idx.list": {
         "section": "llama-index",
         "type": "dict",
@@ -2406,6 +2444,18 @@
         "step": null,
         "advanced": false
     },
+    "log.realtime": {
+        "section": "debug",
+        "type": "bool",
+        "slider": false,
+        "label": "Log Realtime sessions to console",
+        "value": false,
+        "min": null,
+        "max": null,
+        "multiplier": null,
+        "step": null,
+        "advanced": false
+    },
     "log.assistants": {
         "section": "debug",
         "type": "bool",

pygpt_net/data/locale/locale.de.ini

@@ -845,7 +845,7 @@ mode.agent_openai.tooltip = Fortgeschrittene Agenten (OpenAI)
 mode.agent.tooltip = Einfache Agenten (legacy)
 mode.assistant = Assistent
 mode.assistant.tooltip = Chat mittels Assistants API
-mode.audio = Chat mit Audio
+mode.audio = Realtime + audio
 mode.chat = Chat
 mode.chat.tooltip = Chatmodus (Standard)
 mode.completion = Vervollständigung
@@ -1197,9 +1197,9 @@ settings.frequency_penalty = Frequenzstrafe
 settings.func_call.native = Native API-Funktionsaufrufe verwenden
 settings.func_call.native.desc = Wenn aktiviert, benutzt die Anwendung native API-Funktionsaufrufe anstelle des internen pygpt-Formats und der unten stehenden Befehlsprompten. Nur Chat- und Assistenz-Modi.
 settings.img_dialog_open = Bild-Dialog nach Generierung öffnen (Bildmodus)
-settings.img_prompt_model = DALL-E: Modell zur Prompterzeugung
-settings.img_quality = DALL-E: image quality
-settings.img_resolution = DALL-E: Bildgröße
+settings.img_prompt_model = Modell zur Prompterzeugung
+settings.img_quality = image quality
+settings.img_resolution = Bildgröße
 settings.layout.animation.disable = Animationen deaktivieren
 settings.layout.animation.disable.desc = Deaktiviert Layout-Animationen, wie animierte Ladegeräte usw.
 settings.layout.density = Layoutdichte
@@ -1294,7 +1294,7 @@ settings.prompt.ctx.auto_summary.user = Kontext: Auto-Zusammenfassung (Benutzern
 settings.prompt.ctx.auto_summary.user.desc = Platzhalter: {input}, {output}
 settings.prompt.expert = Experte: Masteraufforderung
 settings.prompt.expert.desc = Anweisung (Systemaufforderung) für den Master-Experten, wie man Sklavenexperten handhabt. Anweisungen für Sklavenexperten werden aus ihren Voreinstellungen gegeben.
-settings.prompt.img = DALL-E: Bildgenerierung
+settings.prompt.img = Bildgenerierung
 settings.prompt.img.desc = Aufforderung zur Erzeugung von Anweisungen für DALL-E (falls Rohmodus deaktiviert ist). Nur im Bildmodus.
 settings.remote_tools.code_interpreter = Code-Interpreter
 settings.remote_tools.code_interpreter.desc = Aktivieren Sie das `code_interpreter` Remote-Tool im Chat-Modus / über OpenAI Responses API.

pygpt_net/data/locale/locale.en.ini

@@ -188,6 +188,7 @@ attachments_uploaded.clear.confirm = WARNING: are you sure you want to delete al
 attachments_uploaded.delete.confirm = WARNING: are you sure you want to delete this file from the remote server?
 attachments_uploaded.sync.tip = Tip: click on 'Sync' to retrieve the file list from OpenAI
 attachments_uploaded.tab = Uploaded
+audio.auto_turn = Auto (VAD)
 audio.cache.clear.confirm = Are you sure you want to delete all cached audio files?
 audio.cache.clear.success = OK. All audio cache files cleared.
 audio.control.btn = Voice control
@@ -845,7 +846,7 @@ mode.agent_openai.tooltip = Advanced agents (OpenAI)
 mode.agent.tooltip = Simple agents (legacy)
 mode.assistant = Assistants
 mode.assistant.tooltip = Chat using Assistants API
-mode.audio = Chat with Audio
+mode.audio = Realtime + audio
 mode.chat = Chat
 mode.chat.tooltip = Chat mode (default)
 mode.completion = Completion
@@ -1219,9 +1220,9 @@ settings.frequency_penalty = Frequency Penalty
 settings.func_call.native = Use native API function calls
 settings.func_call.native.desc = If enabled, the application will use native API function calls instead of the internal pygpt format and the command prompts from below will not be used. Chat and Assistants modes ONLY.
 settings.img_dialog_open = Open image dialog after generation (Image mode)
-settings.img_prompt_model = DALL-E: prompt generation model
-settings.img_quality = DALL-E: image quality
-settings.img_resolution = DALL-E: image size
+settings.img_prompt_model = Prompt generation model
+settings.img_quality = Image quality
+settings.img_resolution = Image size
 settings.layout.animation.disable = Disable animations
 settings.layout.animation.disable.desc = Disables layout animations, like animated loaders, etc.
 settings.layout.density = Layout density
@@ -1318,26 +1319,28 @@ settings.prompt.ctx.auto_summary.user = Context: auto-summary (user message)
 settings.prompt.ctx.auto_summary.user.desc = Placeholders: {input}, {output}
 settings.prompt.expert = Expert: Master prompt
 settings.prompt.expert.desc = Instruction (system prompt) for Master expert on how to handle slave experts. Instructions for slave experts are given from their presets.
-settings.prompt.img = DALL-E: image generation
+settings.prompt.img = Image generation
 settings.prompt.img.desc = Prompt for generating prompts for DALL-E (if raw-mode is disabled). Image mode only.
 settings.remote_tools.code_interpreter = Code Interpreter
-settings.remote_tools.code_interpreter.desc = Enable `code_interpreter` remote tool in Chat mode / via OpenAI Responses API.
+settings.remote_tools.code_interpreter.desc = Enable `code_interpreter` remote tool - Responses API only.
 settings.remote_tools.file_search = File search
 settings.remote_tools.file_search.args = File search vector store IDs
 settings.remote_tools.file_search.args.desc = Vector store IDs, separated by comma (,)
-settings.remote_tools.file_search.desc = Enable `file_search` remote tool in Chat mode / via OpenAI Responses API.
+settings.remote_tools.file_search.desc = Enable `file_search` remote tool - Responses API only.
 settings.remote_tools.google.code_interpreter = Code Interpreter
-settings.remote_tools.google.code_interpreter.desc = Enable Code Interpreter remote tool in Chat mode.
-settings.remote_tools.google.web_search = Google Web Search
-settings.remote_tools.google.web_search.desc = Enable Google Search remote tool in Chat mode.
+settings.remote_tools.google.code_interpreter.desc = Enable Code Interpreter remote tool.
+settings.remote_tools.google.url_ctx = URL Context
+settings.remote_tools.google.url_ctx.desc = Enable URL Context remote tool.
+settings.remote_tools.google.web_search = Google Search
+settings.remote_tools.google.web_search.desc = Enable Google Search remote tool.
 settings.remote_tools.image = Image generation
-settings.remote_tools.image.desc = Enable `image_generation` remote tool in Chat mode / via OpenAI Responses API.
+settings.remote_tools.image.desc = Enable `image_generation` remote tool - Responses API only.
 settings.remote_tools.mcp = Remote MCP
 settings.remote_tools.mcp.args = Remote MCP configuration
 settings.remote_tools.mcp.args.desc = Configuration in JSON format (will be used in request)
-settings.remote_tools.mcp.desc = Enable `mcp` remote tool in Chat mode / via OpenAI Responses API.
+settings.remote_tools.mcp.desc = Enable `mcp` remote tool - Responses API only.
 settings.remote_tools.web_search = Web Search
-settings.remote_tools.web_search.desc = Enable `web_search` remote tool in Chat mode / via OpenAI Responses API.
+settings.remote_tools.web_search.desc = Enable `web_search` remote tool - Responses API only.
 settings.render.code_syntax = Code syntax highlight
 settings.render.engine = Rendering engine
 settings.render.open_gl = OpenGL hardware acceleration
@@ -1572,3 +1575,6 @@ vision.capture.manual.captured.success = Image captured from the camera:
 vision.capture.name.prefix = Camera capture:
 vision.capture.options.title = Video capture
 vision.checkbox.tooltip = If checked, the vision model is active. It will be automatically activated upon image upload. You can deactivate it in real-time.
+
+settings.audio.input.vad.prefix = VAD prefix padding (in ms)
+settings.audio.input.vad.silence = VAD end silence (in ms)

pygpt_net/data/locale/locale.es.ini

@@ -846,7 +846,7 @@ mode.agent_openai.tooltip = Agentes avanzados (OpenAI)
 mode.agent.tooltip = Agentes simples (legacy)
 mode.assistant = Asistente
 mode.assistant.tooltip = Chatear usando la API de Asistentes
-mode.audio = Chat con audio
+mode.audio = Realtime + audio
 mode.chat = Chat
 mode.chat.tooltip = Modo de chat (predeterminado)
 mode.completion = Finalización
@@ -1198,9 +1198,9 @@ settings.frequency_penalty = Penalización de frecuencia
 settings.func_call.native = Usar llamadas a funciones API nativas
 settings.func_call.native.desc = Si está habilitado, la aplicación usará llamadas a funciones API nativas en lugar del formato pygpt interno y los prompt de comandos a continuación no se usarán. Solo modos de Chat y Asistentes.
 settings.img_dialog_open = Abrir diálogo de imagen después de generar (Modo imagen)
-settings.img_prompt_model = DALL-E: modelo de generación de indicaciones
-settings.img_quality = DALL-E: calidad de imagen
-settings.img_resolution = DALL-E: tamaño de imagen
+settings.img_prompt_model = Modelo de generación de indicaciones
+settings.img_quality = Calidad de imagen
+settings.img_resolution = Tamaño de imagen
 settings.layout.animation.disable = Desactivar animaciones
 settings.layout.animation.disable.desc = Desactivar animaciones de diseño, como cargadores animados, etc.
 settings.layout.density = Densidad de la disposición
@@ -1295,7 +1295,7 @@ settings.prompt.ctx.auto_summary.user = Contexto: resumen automático (mensaje d
 settings.prompt.ctx.auto_summary.user.desc = Marcadores de posición: {input}, {output}
 settings.prompt.expert = Experto: Master prompt
 settings.prompt.expert.desc = Instrucción (prompt del sistema) para el experto Master cómo manejar a los expertos subordinados. Las instrucciones para los expertos subordinados se dan desde sus presets.
-settings.prompt.img = DALL-E: generación de imagen
+settings.prompt.img = Generación de imagen
 settings.prompt.img.desc = Mensaje para generar comandos para DALL-E (si el modo crudo está desactivado). Solo modo de imagen.
 settings.remote_tools.code_interpreter = Intérprete de Código
 settings.remote_tools.code_interpreter.desc = Habilitar herramienta remota `code_interpreter` en modo Chat / vía API de Respuestas de OpenAI.

pygpt_net/data/locale/locale.fr.ini

@@ -845,7 +845,7 @@ mode.agent_openai.tooltip = Agents avancés (OpenAI)
 mode.agent.tooltip = Agents simples (legacy)
 mode.assistant = Assistant
 mode.assistant.tooltip = Discuter via l'API des Assistants
-mode.audio = Chat avec audio
+mode.audio = Realtime + audio
 mode.chat = Chat
 mode.chat.tooltip = Mode chat (par défaut)
 mode.completion = Complétion
@@ -1197,9 +1197,9 @@ settings.frequency_penalty = Pénalité de fréquence
 settings.func_call.native = Utiliser les appels de fonction API natives
 settings.func_call.native.desc = Si activé, l'application utilisera les appels de fonction API natives au lieu du format interne de pygpt et les commandes prompt ci-dessous ne seront pas utilisées. Modes uniquement Chat et Assistants.
 settings.img_dialog_open = Ouvrir la boîte de dialogue d'image après la génération (Mode image)
-settings.img_prompt_model = DALL-E : modèle de génération d'invite
-settings.img_quality = DALL-E: qualité d'image
-settings.img_resolution = DALL-E : taille de l'image
+settings.img_prompt_model = Modèle de génération d'invite
+settings.img_quality = Qualité d'image
+settings.img_resolution = Taille de l'image
 settings.layout.animation.disable = Désactiver les animations
 settings.layout.animation.disable.desc = Désactive les animations de mise en page, comme les chargeurs animés, etc.
 settings.layout.density = Densité de la disposition
@@ -1294,7 +1294,7 @@ settings.prompt.ctx.auto_summary.user = Contexte: résumé automatique (message
 settings.prompt.ctx.auto_summary.user.desc = Espaces réservés: {input}, {output}
 settings.prompt.expert = Expert : Master prompt
 settings.prompt.expert.desc = Instruction (prompt système) pour l'expert Master sur comment gérer les experts esclaves. Les instructions pour les experts esclaves sont données à partir de leurs presets.
-settings.prompt.img = DALL-E: génération d'image
+settings.prompt.img = Génération d'image
 settings.prompt.img.desc = Prompt pour générer des commandes pour DALL-E (si le mode brut est désactivé). Mode image uniquement.
 settings.remote_tools.code_interpreter = Interpréteur de code
 settings.remote_tools.code_interpreter.desc = Activer l'outil distant `code_interpreter` en mode Chat/ via OpenAI Responses API.

pygpt_net/data/locale/locale.it.ini

@@ -845,7 +845,7 @@ mode.agent_openai.tooltip = Agenti avanzati (OpenAI)
 mode.agent.tooltip = Agenti semplici (legacy)
 mode.assistant = Assistente
 mode.assistant.tooltip = Chattare utilizzando l'API degli Assistenti
-mode.audio = Chat con audio
+mode.audio = Realtime + audio
 mode.chat = Chat
 mode.chat.tooltip = Modalità chat (predefinita)
 mode.completion = Completamento
@@ -1197,9 +1197,9 @@ settings.frequency_penalty = Penale di frequenza
 settings.func_call.native = Usa chiamate di funzione API native
 settings.func_call.native.desc = Se abilitato, l'applicazione utilizzerà le chiamate di funzione API native invece del formato interno pygpt e i prompt di comando di seguito non saranno utilizzati. Solo modalità chat e assistenti.
 settings.img_dialog_open = Apri la finestra di dialogo dell'immagine dopo la generazione (Modalità immagine)
-settings.img_prompt_model = DALL-E: modello di generazione del prompt
-settings.img_quality = DALL-E: qualità dell'immagine
-settings.img_resolution = DALL-E: dimensione dell'immagine
+settings.img_prompt_model = Modello di generazione del prompt
+settings.img_quality = Qualità dell'immagine
+settings.img_resolution = Dimensione dell'immagine
 settings.layout.animation.disable = Disabilita animazioni
 settings.layout.animation.disable.desc = Disabilita le animazioni del layout, come i caricamenti animati, ecc.
 settings.layout.density = Densità del layout
@@ -1294,7 +1294,7 @@ settings.prompt.ctx.auto_summary.user = Contesto: auto-riassunto (messaggio dell
 settings.prompt.ctx.auto_summary.user.desc = Placeholder: {input}, {output}
 settings.prompt.expert = Esperto: Master prompt
 settings.prompt.expert.desc = Istruzione (prompt del sistema) per l'esperto Master su come gestire gli esperti subalterni. Le istruzioni per gli esperti subalterni sono date dalle loro preimpostazioni.
-settings.prompt.img = DALL-E: generazione immagine
+settings.prompt.img = Generazione immagine
 settings.prompt.img.desc = Prompt per generare comandi per DALL-E (se la modalità grezza è disabilitata). Solo modalità immagine.
 settings.remote_tools.code_interpreter = Interprete del codice
 settings.remote_tools.code_interpreter.desc = Abilita l'attrezzo remoto `code_interpreter` in modalità Chat / tramite API delle Risposte di OpenAI.

pygpt_net/data/locale/locale.pl.ini

@@ -846,7 +846,7 @@ mode.agent_openai.tooltip = Zaawansowani agenci (OpenAI)
 mode.agent.tooltip = Prości agenci (legacy)
 mode.assistant = Asystent
 mode.assistant.tooltip = Czat przy użyciu API Asystentów
-mode.audio = Czat Audio
+mode.audio = Realtime + audio
 mode.chat = Czat
 mode.chat.tooltip = Tryb czatu (domyślny)
 mode.completion = Uzupełnianie
@@ -1198,9 +1198,9 @@ settings.frequency_penalty = Frequency Penalty
 settings.func_call.native = Używaj natywnych wywołań funkcji API
 settings.func_call.native.desc = Jeśli włączone, aplikacja będzie używać natywnych wywołań funkcji API zamiast wewnętrznego formatu pygpt i poniższych promptów poleceń. Tylko tryby czatu i asystentów.
 settings.img_dialog_open = Otwórz okno dialogowe obrazu po wygenerowaniu (Tryb obrazu)
-settings.img_prompt_model = DALL-E: model do generowania promptów
-settings.img_quality = DALL-E: jakość obrazu
-settings.img_resolution = DALL-E: rozmiar obrazu
+settings.img_prompt_model = Model do generowania promptów
+settings.img_quality = Jakość obrazu
+settings.img_resolution = Rozmiar obrazu
 settings.layout.animation.disable = Wyłącz animacje
 settings.layout.animation.disable.desc = Wyłącza animacje układu, jak animowane ładowarki, itp.
 settings.layout.density = Rozmiar layoutu
@@ -1295,7 +1295,7 @@ settings.prompt.ctx.auto_summary.user = Kontekst: auto-podsumowanie (wiadomość
 settings.prompt.ctx.auto_summary.user.desc = Placeholdery: {input}, {output}
 settings.prompt.expert = Ekspert: Główna wskazówka
 settings.prompt.expert.desc = Instrukcja (systemowa wskazówka) dla głównego eksperta, jak obsługiwać ekspertów pomocniczych. Instrukcje dla ekspertów pomocniczych są podawane z ich ustawień.
-settings.prompt.img = DALL-E: generowanie obrazu
+settings.prompt.img = Generowanie obrazu
 settings.prompt.img.desc = Prompt do generowania poleceń dla DALL-E (jeśli surowy tryb jest wyłączony). Tylko tryb obrazu.
 settings.remote_tools.code_interpreter = Interpreter kodu
 settings.remote_tools.code_interpreter.desc = Włącz narzędzie `code_interpreter` w trybie Czat / za pośrednictwem OpenAI Responses API.

pygpt_net/data/locale/locale.uk.ini

@@ -845,7 +845,7 @@ mode.agent_openai.tooltip = Просунуті агенти (OpenAI)
 mode.agent.tooltip = Прості агенти (legacy)
 mode.assistant = Помічник
 mode.assistant.tooltip = Чат за допомогою API Асистентів
-mode.audio = Чат з аудіо
+mode.audio = Realtime + audio
 mode.chat = Чат
 mode.chat.tooltip = Режим чату (за замовчуванням)
 mode.completion = Завершення
@@ -1197,9 +1197,9 @@ settings.frequency_penalty = Частотний штраф
 settings.func_call.native = Використовувати рідні виклики функцій API
 settings.func_call.native.desc = Якщо увімкнено, програма буде використовувати рідні виклики функцій API замість внутрішнього формату pygpt і нижченаведених запитів команд не використовуватимуться. Лише режими чату та асистентів.
 settings.img_dialog_open = Відкрити діалогове вікно зображення після генерації (Режим зображення)
-settings.img_prompt_model = DALL-E: модель генерації запиту
-settings.img_quality = DALL-E: якість зображення
-settings.img_resolution = DALL-E: розмір зображення
+settings.img_prompt_model = Модель Генерації Запиту
+settings.img_quality = Якість Зображення
+settings.img_resolution = Розмір Зображення
 settings.layout.animation.disable = Вимкнути анімації
 settings.layout.animation.disable.desc = Вимикає анімації макета, як анімовані завантажувачі тощо.
 settings.layout.density = Щільність компонування
@@ -1294,7 +1294,7 @@ settings.prompt.ctx.auto_summary.user = Контекст: авто-резюме
 settings.prompt.ctx.auto_summary.user.desc = Заповнювачі: {input}, {output}
 settings.prompt.expert = Експерт: Основний запит
 settings.prompt.expert.desc = Інструкція (системний запит) для ведучого експерта, як керувати підеекспертами. Інструкції для підеекспертів даються з їхніх налаштувань.
-settings.prompt.img = DALL-E: генерація зображення
+settings.prompt.img = Генерація зображення
 settings.prompt.img.desc = Підказка для генерації команддля DALL-E (якщо вимкнено сирівний режим). Тільки режим зображення.
 settings.remote_tools.code_interpreter = Інтерпретатор коду
 settings.remote_tools.code_interpreter.desc = Увімкніть віддалений інструмент `code_interpreter` у режимі Чат / через Responses API OpenAI.

pygpt_net/data/locale/locale.zh.ini

@@ -845,7 +845,7 @@ mode.agent_openai.tooltip = 高级代理 (OpenAI)
 mode.agent.tooltip = 简单代理（自主）
 mode.assistant = 助手
 mode.assistant.tooltip = 使用助手API進行聊天
-mode.audio = 语音聊天
+mode.audio = Realtime + audio
 mode.chat = 聊天模式
 mode.chat.tooltip = 聊天模式（預設）
 mode.completion = 完成模式

pygpt_net/data/locale/plugin.audio_input.en.ini

@@ -17,6 +17,10 @@ google_args.tooltip = Provide additional keyword arguments for recognize_google(
 google_cloud_args.description = Additional keyword arguments for r.recognize_google_cloud(audio, **kwargs).
 google_cloud_args.label = Additional keyword arguments
 google_cloud_args.tooltip = Provide additional keyword arguments for recognize_google_cloud()
+google_genai_audio_model.description = Specify Gemini model supporting audio, e.g., gemini-2.5-flash
+google_genai_audio_model.label = Model
+google_genai_audio_prompt.description = System prompt for transcription
+google_genai_audio_prompt.label = System Prompt
 magic_word.description = Activate listening only after the magic word is provided, like 'Hey GPT' or 'OK GPT'. Default: False.
 magic_word.label = Magic word
 magic_word_phrase_length.description = Magic word phrase length. Default: 2.