pygpt-net 2.6.30__py3-none-any.whl → 2.6.32__py3-none-any.whl

pygpt_net/CHANGELOG.txt

@@ -1,3 +1,18 @@
+2.6.32 (2025-09-02)
+
+- Added video generation and support for Google Veo 3 models.
+- Introduced new predefined models: veo-3.0-generate-preview and veo-3.0-fast-generate-preview.
+- Integrated DuckDuckGo as a search provider in the WebSearch plugin.
+- Added "Loop" mode to Realtime + audio mode for automatic turn handling and continuous conversation without manually enabling the microphone.
+
+2.6.31 (2025-09-01)
+
+- Chat with Audio mode renamed to Realtime + audio.
+- Added support for real-time audio models from OpenAI (Realtime API) and Google (Live API), featuring real-time audio integration (beta).
+- Introduced new predefined models: gpt-realtime, gpt-4o-realtime-preview, and gemini-2.5-flash-preview-native-audio-dialog.
+- Included Google Gen AI audio input and output providers in the Audio Input/Output plugins.
+- Added URL Context remote tool support in Google Gen AI.
+
 2.6.30 (2025-08-29)
 
 - Added native Google GenAI API support (beta); live audio is not supported yet (#132).

pygpt_net/__init__.py

@@ -6,15 +6,15 @@
 # GitHub:  https://github.com/szczyglis-dev/py-gpt   #
 # MIT License                                        #
 # Created By  : Marcin Szczygliński                  #
-# Updated Date: 2025.08.29 00:00:00                  #
+# Updated Date: 2025.09.02 00:00:00                  #
 # ================================================== #
 
 __author__ = "Marcin Szczygliński"
 __copyright__ = "Copyright 2025, Marcin Szczygliński"
 __credits__ = ["Marcin Szczygliński"]
 __license__ = "MIT"
-__version__ = "2.6.30"
-__build__ = "2025-08-29"
+__version__ = "2.6.32"
+__build__ = "2025-09-02"
 __maintainer__ = "Marcin Szczygliński"
 __github__ = "https://github.com/szczyglis-dev/py-gpt"
 __report__ = "https://github.com/szczyglis-dev/py-gpt/issues"

pygpt_net/app.py

@@ -6,7 +6,7 @@
 # GitHub:  https://github.com/szczyglis-dev/py-gpt   #
 # MIT License                                        #
 # Created By  : Marcin Szczygliński                  #
-# Updated Date: 2025.08.27 20:00:00                  #
+# Updated Date: 2025.09.02 01:00:00                  #
 # ================================================== #
 
 import os
@@ -171,15 +171,18 @@ from pygpt_net.provider.audio_input.openai_whisper import OpenAIWhisper
 from pygpt_net.provider.audio_input.openai_whisper_local import OpenAIWhisperLocal
 from pygpt_net.provider.audio_input.google_speech_recognition import GoogleSpeechRecognition
 from pygpt_net.provider.audio_input.google_cloud_speech_recognition import GoogleCloudSpeechRecognition
+from pygpt_net.provider.audio_input.google_genai import GoogleGenAIAudioInput
 from pygpt_net.provider.audio_input.bing_speech_recognition import BingSpeechRecognition
 from pygpt_net.provider.audio_output.openai_tts import OpenAITextToSpeech
 from pygpt_net.provider.audio_output.ms_azure_tts import MSAzureTextToSpeech
 from pygpt_net.provider.audio_output.google_tts import GoogleTextToSpeech
+from pygpt_net.provider.audio_output.google_genai_tts import GoogleGenAITextToSpeech
 from pygpt_net.provider.audio_output.eleven_labs import ElevenLabsTextToSpeech
 
 # web search engine providers
 from pygpt_net.provider.web.google_custom_search import GoogleCustomSearch
 from pygpt_net.provider.web.microsoft_bing import MicrosoftBingSearch
+from pygpt_net.provider.web.duckduck_search import DuckDuckGoSearch
 
 # tools
 from pygpt_net.tools.indexer import IndexerTool
@@ -318,10 +321,12 @@ def run(**kwargs):
     launcher.add_audio_input(OpenAIWhisperLocal())
     launcher.add_audio_input(GoogleSpeechRecognition())
     launcher.add_audio_input(GoogleCloudSpeechRecognition())
+    launcher.add_audio_input(GoogleGenAIAudioInput())
     launcher.add_audio_input(BingSpeechRecognition())
     launcher.add_audio_output(OpenAITextToSpeech())
     launcher.add_audio_output(MSAzureTextToSpeech())
     launcher.add_audio_output(GoogleTextToSpeech())
+    launcher.add_audio_output(GoogleGenAITextToSpeech())
     launcher.add_audio_output(ElevenLabsTextToSpeech())
 
     # register custom audio providers
@@ -338,6 +343,7 @@ def run(**kwargs):
     # register web providers
     launcher.add_web(GoogleCustomSearch())
     launcher.add_web(MicrosoftBingSearch())
+    launcher.add_web(DuckDuckGoSearch())
 
     # register custom web providers
     providers = kwargs.get('web', None)

pygpt_net/app_core.py

@@ -6,7 +6,7 @@
 # GitHub:  https://github.com/szczyglis-dev/py-gpt   #
 # MIT License                                        #
 # Created By  : Marcin Szczygliński                  #
-# Updated Date: 2025.08.28 09:00:00                  #
+# Updated Date: 2025.09.01 23:00:00                  #
 # ================================================== #
 
 from pygpt_net.config import Config
@@ -43,6 +43,7 @@ from pygpt_net.core.tabs import Tabs
 from pygpt_net.core.text import Text
 from pygpt_net.core.tokens import Tokens
 from pygpt_net.core.updater import Updater
+from pygpt_net.core.video import Video
 from pygpt_net.core.vision import Vision
 from pygpt_net.core.web import Web
 
@@ -92,6 +93,7 @@ class Core:
         self.text = Text(window)
         self.tokens = Tokens(window)
         self.updater = Updater(window)
+        self.video = Video(window)
         self.vision = Vision(window)
         self.web = Web(window)
 

pygpt_net/config.py

@@ -6,7 +6,7 @@
 # GitHub:  https://github.com/szczyglis-dev/py-gpt   #
 # MIT License                                        #
 # Created By  : Marcin Szczygliński                  #
-# Updated Date: 2025.08.18 01:00:00                  #
+# Updated Date: 2025.09.01 23:00:00                  #
 # ================================================== #
 
 import copy
@@ -66,6 +66,8 @@ class Config:
             "presets": "presets",
             "upload": "upload",
             "tmp": "tmp",
+            "video": "video",
+            "music": "music",
         }
         self._app_path = None
         self._version_cache = version if version else None

pygpt_net/controller/__init__.py

@@ -1,4 +1,4 @@
-# !/usr/bin/env python3
+#!/usr/bin/env python3
 # -*- coding: utf-8 -*-
 # ================================================== #
 # This file is a part of PYGPT package               #
@@ -6,7 +6,7 @@
 # GitHub:  https://github.com/szczyglis-dev/py-gpt   #
 # MIT License                                        #
 # Created By  : Marcin Szczygliński                  #
-# Updated Date: 2025.08.24 23:00:00                  #
+# Updated Date: 2025.09.01 23:00:00                  #
 # ================================================== #
 
 from pygpt_net.controller.access import Access
@@ -29,11 +29,13 @@ from pygpt_net.controller.kernel import Kernel
 from pygpt_net.controller.lang import Lang
 from pygpt_net.controller.launcher import Launcher
 from pygpt_net.controller.layout import Layout
+from pygpt_net.controller.media import Media
 from pygpt_net.controller.mode import Mode
 from pygpt_net.controller.model import Model
 from pygpt_net.controller.notepad import Notepad
 from pygpt_net.controller.painter import Painter
 from pygpt_net.controller.plugins import Plugins
+from pygpt_net.controller.realtime import Realtime
 from pygpt_net.controller.presets import Presets
 from pygpt_net.controller.settings import Settings
 from pygpt_net.controller.theme import Theme
@@ -70,12 +72,14 @@ class Controller:
         self.lang = Lang(window)
         self.launcher = Launcher(window)
         self.layout = Layout(window)
+        self.media = Media(window)
         self.mode = Mode(window)
         self.model = Model(window)
         self.notepad = Notepad(window)
         self.painter = Painter(window)
         self.plugins = Plugins(window)
         self.presets = Presets(window)
+        self.realtime = Realtime(window)
         self.settings = Settings(window)
         self.theme = Theme(window)
         self.tools = Tools(window)
@@ -108,6 +112,8 @@ class Controller:
         self.attachment.setup()
         self.camera.setup_ui()
         self.access.setup()
+        self.realtime.setup()
+        self.media.setup()
 
     def post_setup(self):
         """Post-setup, after plugins are loaded"""
@@ -166,6 +172,7 @@ class Controller:
         self.lang.reload()
         self.debug.reload()
         self.chat.reload()
+        self.media.reload()
         self.window.tools.on_reload()
         self.access.reload()
         self.tools.reload()

pygpt_net/controller/audio/audio.py

@@ -6,7 +6,7 @@
 # GitHub:  https://github.com/szczyglis-dev/py-gpt   #
 # MIT License                                        #
 # Created By  : Marcin Szczygliński                  #
-# Updated Date: 2025.08.27 07:00:00                  #
+# Updated Date: 2025.08.31 23:00:00                  #
 # ================================================== #
 
 import os
@@ -40,9 +40,46 @@ class Audio:
     def setup(self):
         """Setup controller"""
         self.update()
+
+        # continuous input (notepad)
         if self.window.core.config.get("audio.input.continuous", False):
             self.window.ui.plugin_addon['audio.input.btn'].continuous.setChecked(True)
 
+        # auto turn (VAD)
+        if self.window.core.config.get("audio.input.auto_turn", False):
+            self.window.ui.nodes['audio.auto_turn'].box.setChecked(True)
+
+        # loop recording
+        if self.window.core.config.get("audio.input.loop", False):
+            self.window.ui.nodes['audio.loop'].box.setChecked(True)
+
+    def execute_input_stop(self):
+        """Execute input stop (from UI)"""
+        self.window.dispatch(Event(Event.AUDIO_INPUT_RECORD_TOGGLE, {
+            "state": False,
+            "auto": True,  # do not emit manual event
+        }))
+
+    def is_recording(self) -> bool:
+        """
+        Check if audio input is recording
+
+        :return: True if recording
+        """
+        return self.window.core.plugins.get("audio_input").is_recording()
+
+    def toggle_auto_turn(self):
+        """Toggle auto turn setting"""
+        value = self.window.ui.nodes['audio.auto_turn'].box.isChecked()
+        self.window.core.config.set("audio.input.auto_turn", value)
+        self.window.core.config.save()
+
+    def toggle_loop(self):
+        """Toggle loop recording setting"""
+        value = self.window.ui.nodes['audio.loop'].box.isChecked()
+        self.window.core.config.set("audio.input.loop", value)
+        self.window.core.config.save()
+
     def toggle_input(
             self,
             state: bool,

pygpt_net/controller/audio/ui.py

@@ -183,7 +183,7 @@ class UI:
         """
         self.recording = True
         self.window.ui.nodes['input'].set_icon_state("mic", True)
-        if mode == "input":
+        if mode in ["input", "realtime"]:
             self.window.controller.chat.common.lock_input()
             return
         btn = self.get_input_btn() if mode == 'input' else self.get_input_control_btn()
@@ -198,7 +198,7 @@ class UI:
         """
         self.recording = False
         self.window.ui.nodes['input'].set_icon_state("mic", False)
-        if mode == "input":
+        if mode in ["input", "realtime"]:
             self.window.controller.chat.common.unlock_input()
             return
         btn = self.get_input_btn() if mode == 'input' else self.get_input_control_btn()

pygpt_net/controller/chat/audio.py

@@ -6,7 +6,7 @@
 # GitHub:  https://github.com/szczyglis-dev/py-gpt   #
 # MIT License                                        #
 # Created By  : Marcin Szczygliński                  #
-# Updated Date: 2024.12.14 18:00:00                  #
+# Updated Date: 2025.08.30 06:00:00                  #
 # ================================================== #
 
 import base64
@@ -40,19 +40,12 @@ class Audio:
         """Update input/output audio"""
         mode = self.window.core.config.get("mode")
         if mode == MODE_AUDIO:
-            if not self.window.controller.audio.is_output_enabled():
-                self.window.controller.audio.enable_output()
-                self.tmp_output = True
-            else:
-                self.tmp_output = False
             if not self.window.controller.audio.is_input_enabled():
                 self.window.controller.audio.enable_input()
                 self.tmp_input = True
             else:
                 self.tmp_input = False
         else:
-            if self.tmp_output:
-                self.window.controller.audio.disable_output()
             if self.tmp_input:
                 self.window.controller.audio.disable_input()
 

pygpt_net/controller/chat/common.py

@@ -6,16 +6,17 @@
 # GitHub:  https://github.com/szczyglis-dev/py-gpt   #
 # MIT License                                        #
 # Created By  : Marcin Szczygliński                  #
-# Updated Date: 2025.08.27 07:00:00                  #
+# Updated Date: 2025.09.01 23:00:00                  #
 # ================================================== #
 
 import os
+from typing import Any
 
 from PySide6.QtGui import QTextCursor
 from PySide6.QtWidgets import QFileDialog, QApplication
 
 from pygpt_net.core.events import Event, AppEvent, RenderEvent, KernelEvent
-from pygpt_net.core.types import MODE_ASSISTANT
+from pygpt_net.core.types import MODE_ASSISTANT, MODE_AUDIO
 from pygpt_net.item.ctx import CtxItem
 from pygpt_net.item.model import ModelItem
 from pygpt_net.utils import trans
@@ -99,26 +100,6 @@ class Common:
         event = RenderEvent(RenderEvent.ON_SWITCH)
         self.window.dispatch(event)  # switch renderer if needed
 
-        # edit icons
-        """
-        if self.window.core.config.has('ctx.edit_icons'):
-            self.window.ui.nodes['output.edit'].setChecked(self.window.core.config.get('ctx.edit_icons'))
-            data = {
-                "initialized": self.initialized,
-            }
-            if self.window.core.config.get('ctx.edit_icons'):
-                event = RenderEvent(RenderEvent.ON_EDIT_ENABLE, data)
-            else:
-                event = RenderEvent(RenderEvent.ON_EDIT_DISABLE, data)
-            self.window.dispatch(event)
-        """
-
-        # images generation
-        if self.window.core.config.get('img_raw'):
-            self.window.ui.config['global']['img_raw'].setChecked(True)
-        else:
-            self.window.ui.config['global']['img_raw'].setChecked(False)
-
         # set focus to input
         self.window.ui.nodes['input'].setFocus()
         self.initialized = True
@@ -257,7 +238,7 @@ class Common:
             self.window.controller.access.voice.stop_recording(timeout=True)
 
         if self.window.core.plugins.get("audio_input").handler_simple.is_recording:
-            self.window.core.plugins.get("audio_input").handler_simple.stop_recording(timeout=False)
+            self.window.dispatch(Event(Event.AUDIO_INPUT_RECORD_TOGGLE))
             return
 
         # stop audio output if playing
@@ -275,7 +256,8 @@ class Common:
         """
         # don't unlock input and leave stop btn if assistant mode or if agent/autonomous is enabled
         # send btn will be unlocked in agent mode on stop
-        if self.can_unlock(ctx):
+        mode = self.window.core.config.get('mode')
+        if self.can_unlock(ctx) and mode != MODE_AUDIO:
             if not self.window.controller.kernel.stopped():
                 self.unlock_input()  # unlock input
                 return True
@@ -391,6 +373,23 @@ class Common:
             event = RenderEvent(RenderEvent.ON_TS_DISABLE, data)
         self.window.dispatch(event)
 
+    def toggle_edit_icons(self, value: bool):
+        """
+        Toggle edit icons
+
+        :param value: value of the checkbox
+        """
+        self.window.core.config.set('ctx.edit_icons', value)
+        self.window.core.config.save()
+        data = {
+            "initialized": True,
+        }
+        if value:
+            event = RenderEvent(RenderEvent.ON_EDIT_ENABLE, data)
+        else:
+            event = RenderEvent(RenderEvent.ON_EDIT_DISABLE, data)
+        self.window.dispatch(event)
+
     def toggle_raw(self, value: bool):
         """
         Toggle raw (plain) output
@@ -414,44 +413,6 @@ class Common:
         # restore previous font size
         self.window.controller.ui.update_font_size()
 
-    def toggle_edit_icons(self, value: bool):
-        """
-        Toggle edit icons
-
-        :param value: value of the checkbox
-        """
-        self.window.core.config.set('ctx.edit_icons', value)
-        self.window.core.config.save()
-        data = {
-            "initialized": True,
-        }
-        if value:
-            event = RenderEvent(RenderEvent.ON_EDIT_ENABLE, data)
-        else:
-            event = RenderEvent(RenderEvent.ON_EDIT_DISABLE, data)
-        self.window.dispatch(event)
-
-    def img_enable_raw(self):
-        """Enable help for images"""
-        self.window.core.config.set('img_raw', True)
-        self.window.core.config.save()
-
-    def img_disable_raw(self):
-        """Disable help for images"""
-        self.window.core.config.set('img_raw', False)
-        self.window.core.config.save()
-
-    def img_toggle_raw(self, state: bool):
-        """
-        Toggle help for images
-
-        :param state: state of checkbox
-        """
-        if not state:
-            self.img_disable_raw()
-        else:
-            self.img_enable_raw()
-
     def save_text(
             self,
             text: str,