pygpt-net 2.5.17__tar.gz → 2.5.18__tar.gz

{pygpt_net-2.5.17 → pygpt_net-2.5.18}/CHANGELOG.md

@@ -1,5 +1,12 @@
 # CHANGELOG
 
+## 2.5.18 (2025-06-26)
+
+- Non-GPT models are now available in standard Chat mode.
+- Added a new remote tool: `image_generation` in Responses API -> disabled by default, enable in `Config -> Settings -> Remote Tools`. Enables native image generation and editing of uploaded images in Chat mode.
+- Added a new model `gpt-image-1` and improved image generation.
+- Other small fixes.
+
 ## 2.5.17 (2025-06-25)
 
 - Added settings for enable/disable Remote Tools via Responses API in Chat mode: Config -> Settings -> Remote tools. Currently only web-search-preview tool is available, rest of tools coming soon.

{pygpt_net-2.5.17 → pygpt_net-2.5.18}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.3
 Name: pygpt-net
-Version: 2.5.17
+Version: 2.5.18
 Summary: Desktop AI Assistant powered by models: OpenAI o1, GPT-4o, GPT-4, GPT-4 Vision, GPT-3.5, DALL-E 3, Llama 3, Mistral, Gemini, Claude, DeepSeek, Bielik, and other models supported by Langchain, Llama Index, and Ollama. Features include chatbot, text completion, image generation, vision analysis, speech-to-text, internet access, file handling, command execution and more.
 License: MIT
 Keywords: py_gpt,py-gpt,pygpt,desktop,app,o1,gpt,gpt4,gpt-4o,gpt-4v,gpt3.5,gpt-4,gpt-4-vision,gpt-3.5,llama3,mistral,gemini,deepseek,bielik,claude,tts,whisper,vision,chatgpt,dall-e,chat,chatbot,assistant,text completion,image generation,ai,api,openai,api key,langchain,llama-index,ollama,presets,ui,qt,pyside
@@ -100,7 +100,7 @@ Description-Content-Type: text/markdown
 
 [![pygpt](https://snapcraft.io/pygpt/badge.svg)](https://snapcraft.io/pygpt)
 
-Release: **2.5.17** | build: **2025-06-25** | Python: **>=3.10, <3.13**
+Release: **2.5.18** | build: **2025-06-26** | Python: **>=3.10, <3.13**
 
 > Official website: https://pygpt.net | Documentation: https://pygpt.readthedocs.io
 > 
@@ -462,9 +462,9 @@ Your API keys will be available here:
 
 **+ Inline Vision and Image generation**
 
-This mode in **PyGPT** mirrors `ChatGPT`, allowing you to chat with models such as `o1`, `GPT-4`, `GPT-4o` and `GPT-3.5`. It works by using the `ChatCompletion` OpenAI API.
+This mode in **PyGPT** mirrors `ChatGPT`, allowing you to chat with models such as `o1`, `o3`, `GPT-4`, `GPT-4o` and `GPT-3.5`. It works by using the `Responses` OpenAI API.
 
-**Tip: This mode directly uses the OpenAI API. If you want to use models other than GPT (such as Gemini, Claude, or Llama3), use `Chat with Files` mode.**
+**Tip: This mode directly uses the OpenAI API. Other models, such as Gemini, Claude, or Llama3, are supported in Chat mode via LlamaIndex, which the application switches to in the background when working with models other than OpenAI.**
 
 The main part of the interface is a chat window where you see your conversations. Below it is a message box for typing. On the right side, you can set up or change the model and system prompt. You can also save these settings as presets to easily switch between models or tasks.
 
@@ -520,7 +520,7 @@ From version `2.0.107` the `davinci` models are deprecated and has been replaced
 
 ### DALL-E 3
 
-**PyGPT** enables quick and easy image creation with `DALL-E 3`. 
+**PyGPT** enables quick and easy image creation with `DALL-E 3` or `gpt-image-1`. 
 The older model version, `DALL-E 2`, is also accessible. Generating images is akin to a chat conversation  -  a user's prompt triggers the generation, followed by downloading, saving to the computer, 
 and displaying the image onscreen. You can send raw prompt to `DALL-E` in `Image generation` mode or ask the model for the best prompt.
 
@@ -1095,7 +1095,7 @@ There is built-in support for those LLM providers:
 
 How to use locally installed Llama 3 or Mistral models:
 
-1) Choose a working mode: `Chat with Files`.
+1) Choose a working mode: `Chat` or `Chat with Files`.
 
 2) On the models list - select, edit, or add a new model (with `ollama` provider). You can edit the model settings through the menu `Config -> Models -> Edit`, then configure the model parameters in the `advanced` section.
 
@@ -4124,6 +4124,13 @@ may consume additional tokens that are not displayed in the main window.
 
 ## Recent changes:
 
+**2.5.18 (2025-06-26)**
+
+- Non-GPT models are now available in standard Chat mode.
+- Added a new remote tool: `image_generation` in Responses API -> disabled by default, enable in `Config -> Settings -> Remote Tools`. Enables native image generation and editing of uploaded images in Chat mode.
+- Added a new model `gpt-image-1` and improved image generation.
+- Other small fixes.
+
 **2.5.17 (2025-06-25)**
 
 - Added settings for enable/disable Remote Tools via Responses API in Chat mode: Config -> Settings -> Remote tools. Currently only web-search-preview tool is available, rest of tools coming soon.

{pygpt_net-2.5.17 → pygpt_net-2.5.18}/README.md

@@ -2,7 +2,7 @@
 
 [![pygpt](https://snapcraft.io/pygpt/badge.svg)](https://snapcraft.io/pygpt)
 
-Release: **2.5.17** | build: **2025-06-25** | Python: **>=3.10, <3.13**
+Release: **2.5.18** | build: **2025-06-26** | Python: **>=3.10, <3.13**
 
 > Official website: https://pygpt.net | Documentation: https://pygpt.readthedocs.io
 > 
@@ -364,9 +364,9 @@ Your API keys will be available here:
 
 **+ Inline Vision and Image generation**
 
-This mode in **PyGPT** mirrors `ChatGPT`, allowing you to chat with models such as `o1`, `GPT-4`, `GPT-4o` and `GPT-3.5`. It works by using the `ChatCompletion` OpenAI API.
+This mode in **PyGPT** mirrors `ChatGPT`, allowing you to chat with models such as `o1`, `o3`, `GPT-4`, `GPT-4o` and `GPT-3.5`. It works by using the `Responses` OpenAI API.
 
-**Tip: This mode directly uses the OpenAI API. If you want to use models other than GPT (such as Gemini, Claude, or Llama3), use `Chat with Files` mode.**
+**Tip: This mode directly uses the OpenAI API. Other models, such as Gemini, Claude, or Llama3, are supported in Chat mode via LlamaIndex, which the application switches to in the background when working with models other than OpenAI.**
 
 The main part of the interface is a chat window where you see your conversations. Below it is a message box for typing. On the right side, you can set up or change the model and system prompt. You can also save these settings as presets to easily switch between models or tasks.
 
@@ -422,7 +422,7 @@ From version `2.0.107` the `davinci` models are deprecated and has been replaced
 
 ### DALL-E 3
 
-**PyGPT** enables quick and easy image creation with `DALL-E 3`. 
+**PyGPT** enables quick and easy image creation with `DALL-E 3` or `gpt-image-1`. 
 The older model version, `DALL-E 2`, is also accessible. Generating images is akin to a chat conversation  -  a user's prompt triggers the generation, followed by downloading, saving to the computer, 
 and displaying the image onscreen. You can send raw prompt to `DALL-E` in `Image generation` mode or ask the model for the best prompt.
 
@@ -997,7 +997,7 @@ There is built-in support for those LLM providers:
 
 How to use locally installed Llama 3 or Mistral models:
 
-1) Choose a working mode: `Chat with Files`.
+1) Choose a working mode: `Chat` or `Chat with Files`.
 
 2) On the models list - select, edit, or add a new model (with `ollama` provider). You can edit the model settings through the menu `Config -> Models -> Edit`, then configure the model parameters in the `advanced` section.
 
@@ -4026,6 +4026,13 @@ may consume additional tokens that are not displayed in the main window.
 
 ## Recent changes:
 
+**2.5.18 (2025-06-26)**
+
+- Non-GPT models are now available in standard Chat mode.
+- Added a new remote tool: `image_generation` in Responses API -> disabled by default, enable in `Config -> Settings -> Remote Tools`. Enables native image generation and editing of uploaded images in Chat mode.
+- Added a new model `gpt-image-1` and improved image generation.
+- Other small fixes.
+
 **2.5.17 (2025-06-25)**
 
 - Added settings for enable/disable Remote Tools via Responses API in Chat mode: Config -> Settings -> Remote tools. Currently only web-search-preview tool is available, rest of tools coming soon.

{pygpt_net-2.5.17 → pygpt_net-2.5.18}/pyproject.toml

@@ -1,6 +1,6 @@
 [tool.poetry]
 name = "pygpt-net"
-version = "2.5.17"
+version = "2.5.18"
 description = "Desktop AI Assistant powered by models: OpenAI o1, GPT-4o, GPT-4, GPT-4 Vision, GPT-3.5, DALL-E 3, Llama 3, Mistral, Gemini, Claude, DeepSeek, Bielik, and other models supported by Langchain, Llama Index, and Ollama. Features include chatbot, text completion, image generation, vision analysis, speech-to-text, internet access, file handling, command execution and more."
 authors = ["Marcin Szczyglinski <info@pygpt.net>"]
 license = "MIT"

{pygpt_net-2.5.17 → pygpt_net-2.5.18}/src/pygpt_net/CHANGELOG.txt

@@ -1,3 +1,10 @@
+2.5.18 (2025-06-26)
+
+- Non-GPT models are now available in standard Chat mode.
+- Added a new remote tool: `image_generation` in Responses API -> disabled by default, enable in `Config -> Settings -> Remote Tools`. Enables native image generation and editing of uploaded images in Chat mode.
+- Added a new model `gpt-image-1` and improved image generation.
+- Other small fixes.
+
 2.5.17 (2025-06-25)
 
 - Added settings for enable/disable Remote Tools via Responses API in Chat mode: Config -> Settings -> Remote tools. Currently only web-search-preview tool is available, rest of tools coming soon.

{pygpt_net-2.5.17 → pygpt_net-2.5.18}/src/pygpt_net/__init__.py

@@ -6,15 +6,15 @@
 # GitHub:  https://github.com/szczyglis-dev/py-gpt   #
 # MIT License                                        #
 # Created By  : Marcin Szczygliński                  #
-# Updated Date: 2025-06-25 02:00:00                  #
+# Updated Date: 2025-06-26 02:00:00                  #
 # ================================================== #
 
 __author__ = "Marcin Szczygliński"
 __copyright__ = "Copyright 2025, Marcin Szczygliński"
 __credits__ = ["Marcin Szczygliński"]
 __license__ = "MIT"
-__version__ = "2.5.17"
-__build__ = "2025-06-25"
+__version__ = "2.5.18"
+__build__ = "2025-06-26"
 __maintainer__ = "Marcin Szczygliński"
 __github__ = "https://github.com/szczyglis-dev/py-gpt"
 __report__ = "https://github.com/szczyglis-dev/py-gpt/issues"

{pygpt_net-2.5.17 → pygpt_net-2.5.18}/src/pygpt_net/controller/chat/common.py

@@ -282,15 +282,17 @@ class Common:
         if not exit:
             self.window.dispatch(AppEvent(AppEvent.INPUT_STOPPED))  # app event
 
-    def check_api_key(self) -> bool:
+    def check_api_key(self, monit: bool = True) -> bool:
         """
         Check if API KEY is set
 
+        :param monit: True if monitor should be shown
         :return: True if API KEY is set, False otherwise
         """
         result = True
         if self.window.core.config.get('api_key') is None or self.window.core.config.get('api_key') == '':
-            self.window.controller.launcher.show_api_monit()
+            if monit:
+                self.window.controller.launcher.show_api_monit()
             self.window.update_status("Missing API KEY!")
             result = False
         return result

{pygpt_net-2.5.17 → pygpt_net-2.5.18}/src/pygpt_net/controller/chat/input.py

@@ -6,8 +6,9 @@
 # GitHub:  https://github.com/szczyglis-dev/py-gpt   #
 # MIT License                                        #
 # Created By  : Marcin Szczygliński                  #
-# Updated Date: 2025.06.24 16:00:00                  #
+# Updated Date: 2025.06.26 16:00:00                  #
 # ================================================== #
+
 import os
 from typing import Optional, Any, Dict
 
@@ -20,6 +21,7 @@ from pygpt_net.core.types import (
     MODE_LLAMA_INDEX,
     MODE_ASSISTANT,
     MODE_IMAGE,
+    MODE_CHAT,
 )
 from pygpt_net.item.ctx import CtxItem
 from pygpt_net.utils import trans
@@ -76,27 +78,29 @@ class Input:
 
             # check ollama model
             model = self.window.core.config.get('model')
-            if mode == MODE_LLAMA_INDEX and model is not None:
+            if model:
                 model_data = self.window.core.models.get(model)
                 if model_data is not None and model_data.is_ollama():
-                    model_id = model_data.get_ollama_model()
-                    # load ENV vars first
-                    if ('env' in model_data.llama_index
-                            and model_data.llama_index['env'] is not None):
-                        for item in model_data.llama_index['env']:
-                            key = item.get('name', '').strip()
-                            value = item.get('value', '').strip()
-                            os.environ[key] = value
-                    status = self.window.core.models.ollama.check_model(model_id)
-                    is_installed = status.get('is_installed', False)
-                    is_model = status.get('is_model', False)
-                    if not is_installed:
-                        self.window.ui.dialogs.alert(trans("dialog.ollama.not_installed"))
-                        return
-                    if not is_model:
-                        self.window.ui.dialogs.alert(
-                            trans("dialog.ollama.model_not_found").replace("{model}", model_id))
-                        return
+                    if (mode == MODE_LLAMA_INDEX or
+                            (mode == MODE_CHAT and not model_data.is_openai() and model_data.is_ollama())):
+                        model_id = model_data.get_ollama_model()
+                        # load ENV vars first
+                        if ('env' in model_data.llama_index
+                                and model_data.llama_index['env'] is not None):
+                            for item in model_data.llama_index['env']:
+                                key = item.get('name', '').strip()
+                                value = item.get('value', '').strip()
+                                os.environ[key] = value
+                        status = self.window.core.models.ollama.check_model(model_id)
+                        is_installed = status.get('is_installed', False)
+                        is_model = status.get('is_model', False)
+                        if not is_installed:
+                            self.window.ui.dialogs.alert(trans("dialog.ollama.not_installed"))
+                            return
+                        if not is_model:
+                            self.window.ui.dialogs.alert(
+                                trans("dialog.ollama.model_not_found").replace("{model}", model_id))
+                            return
 
         # listen for stop command
         if self.generating \
@@ -247,14 +251,19 @@ class Input:
             self.generating = False  # unlock as not generating
             return
 
-        # check API key, show monit if no API key
+        # check OpenAI API key, show monit if no API key
         if mode not in self.window.controller.launcher.no_api_key_allowed:
-            if not self.window.controller.chat.common.check_api_key():
-                self.generating = False
-                self.window.dispatch(KernelEvent(KernelEvent.STATE_ERROR, {
-                    "id": "chat",
-                }))
-                return
+            if not self.window.controller.chat.common.check_api_key(monit=False):
+                model = self.window.core.config.get('model')
+                if model:
+                    model_data = self.window.core.models.get(model)
+                    if model_data is not None and model_data.is_openai():
+                        self.window.controller.chat.common.check_api_key(monit=True)
+                        self.generating = False
+                        self.window.dispatch(KernelEvent(KernelEvent.STATE_ERROR, {
+                            "id": "chat",
+                        }))
+                        return
 
         # set state to: busy
         self.window.dispatch(KernelEvent(KernelEvent.STATE_BUSY, {

{pygpt_net-2.5.17 → pygpt_net-2.5.18}/src/pygpt_net/controller/chat/stream.py

@@ -8,7 +8,8 @@
 # Created By  : Marcin Szczygliński                  #
 # Updated Date: 2025.06.25 02:00:00                  #
 # ================================================== #
-
+import base64
+import uuid
 from typing import Any
 
 from PySide6.QtWidgets import QApplication
@@ -39,6 +40,8 @@ class Stream:
         tool_calls = []
         fn_args_buffers = {}
         citations = []
+        img_path = self.window.core.image.gen_unique_path(ctx)
+        is_image = False
 
         # chunks: stream begin
         data = {
@@ -121,8 +124,8 @@ class Stream:
                                 if tool_chunk.function.arguments:
                                     tool_call["function"]["arguments"] += tool_chunk.function.arguments
 
+                    # OpenAI Responses API
                     elif chunk_type == "api_chat_responses":
-
                         if etype == "response.output_text.delta":
                             response = chunk.delta
 
@@ -154,6 +157,19 @@ class Stream:
                                 citations.append(url_citation)
                                 ctx.urls = citations
 
+                        # ---------- image gen ----------
+                        elif etype == "response.image_generation_call.partial_image":
+                            idx = chunk.partial_image_index
+                            image_base64 = chunk.partial_image_b64
+                            image_bytes = base64.b64decode(image_base64)
+                            with open(img_path, "wb") as f:
+                                f.write(image_bytes)
+                            is_image = True
+
+                        # ---------- response ID ----------
+                        elif etype == "response.created":
+                            ctx.msg_id = chunk.response.id # store previous response ID
+
                         # ---------- end / error ----------
                         elif etype in {"response.done", "response.failed", "error"}:
                             pass
@@ -214,6 +230,10 @@ class Stream:
                 if tool_calls:
                     self.window.core.command.unpack_tool_calls_chunks(ctx, tool_calls)
 
+                # append images
+                if is_image:
+                    ctx.images = [img_path]  # save image path to ctx
+
         except Exception as e:
             self.window.core.debug.log(e)
             error = e

{pygpt_net-2.5.17 → pygpt_net-2.5.18}/src/pygpt_net/controller/config/placeholder.py

@@ -270,7 +270,7 @@ class Placeholder:
             model = models[id]
             suffix = ""
             if "provider" in model.llama_index and model.llama_index["provider"] == "ollama":
-                suffix = " [Ollama]"
+                suffix = " (Ollama)"
             name = model.name + suffix
             data.append({id: name})
         return data

{pygpt_net-2.5.17 → pygpt_net-2.5.18}/src/pygpt_net/controller/model/__init__.py

@@ -161,7 +161,7 @@ class Model:
         for k in data:
             suffix = ""
             if "provider" in data[k].llama_index and data[k].llama_index["provider"] == "ollama":
-                suffix = " [Ollama]"
+                suffix = " (Ollama)"
             items[k] = data[k].name + suffix
         items = dict(sorted(items.items(), key=lambda item: item[1]))  # sort items by name
         self.window.ui.nodes["prompt.model"].set_keys(items)

{pygpt_net-2.5.17 → pygpt_net-2.5.18}/src/pygpt_net/controller/model/editor.py

@@ -6,7 +6,7 @@
 # GitHub:  https://github.com/szczyglis-dev/py-gpt   #
 # MIT License                                        #
 # Created By  : Marcin Szczygliński                  #
-# Updated Date: 2024.12.14 08:00:00                  #
+# Updated Date: 2025.06.26 16:00:00                  #
 # ================================================== #
 
 import copy
@@ -56,6 +56,11 @@ class Editor:
                 "type": "bool",
                 "label": "model.default",
             },
+            "openai": {
+                "type": "bool",
+                "label": "model.openai",
+                "description": "model.openai.desc",
+            },
             "langchain.provider": {
                 "type": "combo",
                 "use": "langchain_providers",

{pygpt_net-2.5.17 → pygpt_net-2.5.18}/src/pygpt_net/controller/model/importer.py

@@ -6,7 +6,7 @@
 # GitHub:  https://github.com/szczyglis-dev/py-gpt   #
 # MIT License                                        #
 # Created By  : Marcin Szczygliński                  #
-# Updated Date: 2025.06.24 02:00:00                  #
+# Updated Date: 2025.06.26 16:00:00                  #
 # ================================================== #
 
 import copy
@@ -200,7 +200,7 @@ class Importer:
         """
         Get current ollama models
 
-        :return: ollama models dictionary
+        :return: PyGPT ollama models dictionary
         """
         items = copy.deepcopy(self.window.core.models.items)
         for key in list(items.keys()):
@@ -214,7 +214,7 @@ class Importer:
         """
         Get available ollama models
 
-        :return: ollama models dictionary
+        :return: Ollama API models dictionary
         """
         models = {}
         status = self.window.core.models.ollama.get_status()
@@ -233,6 +233,7 @@ class Importer:
                     m.id = name
                     m.name = name
                     m.mode = [
+                        "chat",
                         "llama_index",
                         "agent",
                         "agent_llama",

{pygpt_net-2.5.17 → pygpt_net-2.5.18}/src/pygpt_net/core/bridge/__init__.py

@@ -6,7 +6,7 @@
 # GitHub:  https://github.com/szczyglis-dev/py-gpt   #
 # MIT License                                        #
 # Created By  : Marcin Szczygliński                  #
-# Updated Date: 2025.03.02 19:00:00                  #
+# Updated Date: 2025.06.26 16:00:00                  #
 # ================================================== #
 
 import time
@@ -93,9 +93,11 @@ class Bridge:
         if model is not None:
             if not model.is_supported(mode):  # check selected mode
                 mode = self.window.core.models.get_supported_mode(model, mode)  # switch
+                if base_mode == MODE_CHAT and mode == MODE_LLAMA_INDEX:
+                    context.idx = None # disable index if in Chat mode and switch to Llama Index
 
         if mode == MODE_LLAMA_INDEX and base_mode != MODE_LLAMA_INDEX:
-            context.idx_mode = MODE_CHAT
+            context.idx_mode = MODE_CHAT  # default in sub-mode
 
         if is_virtual:
             if mode == MODE_LLAMA_INDEX:  # after switch
@@ -192,11 +194,13 @@ class Bridge:
                 debug = {k: str(v) for k, v in context.to_dict().items()}
                 self.window.core.debug.debug(str(debug))
 
+        # --- DEBUG ONLY ---
         self.last_context_quick = context  # store last context for quick call (debug)
 
         if context.model is not None:
-            # check if model is supported by chat API, if not then try to use llama-index or langchain call
-            if not context.model.is_supported(MODE_CHAT) and not context.model.is_supported(MODE_RESEARCH):
+            # check if model is supported by OpenAI API, if not then try to use llama-index or langchain call
+            if ((not context.model.is_supported(MODE_CHAT) or not context.model.is_openai())
+                    and not context.model.is_supported(MODE_RESEARCH)):
 
                 # tmp switch to: llama-index
                 if context.model.is_supported(MODE_LLAMA_INDEX):

{pygpt_net-2.5.17 → pygpt_net-2.5.18}/src/pygpt_net/core/command/__init__.py

@@ -241,6 +241,8 @@ class Command:
         parsed = []
         for tool_call in tool_calls:
             try:
+                if not hasattr(tool_call, 'name'):
+                    continue
                 parsed.append(
                     {
                         "id": tool_call.id,
@@ -615,7 +617,7 @@ class Command:
         :return: True if enabled
         """
         disabled_modes = [
-            MODE_LLAMA_INDEX,
+            #MODE_LLAMA_INDEX,
             MODE_LANGCHAIN,
             MODE_COMPLETION,
         ]
@@ -624,6 +626,13 @@ class Command:
             return False  # disabled for specific modes
         if self.window.controller.agent.legacy.enabled() or self.window.controller.agent.experts.enabled():
             return False
+        model = self.window.core.config.get('model')
+        if model:
+            model_data = self.window.core.models.get(model)
+            if model_data:
+                llama_provider = model_data.get_llama_provider()
+                if llama_provider in self.window.core.idx.chat.tool_calls_not_allowed_providers:
+                    return False
         return self.window.core.config.get('func_call.native', False)  # otherwise check config
 
     def is_enabled(self, cmd: str) -> bool:

{pygpt_net-2.5.17 → pygpt_net-2.5.18}/src/pygpt_net/core/idx/chat.py

@@ -6,7 +6,7 @@
 # GitHub:  https://github.com/szczyglis-dev/py-gpt   #
 # MIT License                                        #
 # Created By  : Marcin Szczygliński                  #
-# Updated Date: 2025.06.21 23:00:00                  #
+# Updated Date: 2025.06.26 16:00:00                  #
 # ================================================== #
 
 import json
@@ -39,6 +39,7 @@ class Chat:
         self.context = Context(window)
         self.tool_calls_not_allowed_providers = [
             "ollama",
+            "google",
             "hugging_face_api",
             "deepseek_api",
         ]
@@ -237,6 +238,7 @@ class Chat:
         use_index = True
         verbose = self.window.core.config.get("log.llama", False)
         allow_native_tool_calls = True
+        response = None
         if ('provider' in model.llama_index
                 and model.llama_index['provider'] in self.tool_calls_not_allowed_providers):
             allow_native_tool_calls = False
@@ -246,10 +248,12 @@ class Chat:
             use_index = False
 
         # disable index if no api key
+        """
         if self.window.core.config.get("api_key") == "" and self.window.core.config.get("llama.idx.embeddings.provider") == "openai":
             print("Warning: no api key! Disabling index...")
             chat_mode = "simple"  # do not use query engine if no index
             use_index = False
+        """
 
         if model is None or not isinstance(model, ModelItem):
             raise Exception("Model config not provided")
@@ -322,6 +326,7 @@ class Chat:
             if stream:
                 # IMPORTANT: stream chat with tools not supported by all providers
                 if allow_native_tool_calls and hasattr(llm, "stream_chat_with_tools"):
+                    self.log("Using with tools...")
                     response = llm.stream_chat_with_tools(
                         tools=tools,
                         messages=history,

{pygpt_net-2.5.17 → pygpt_net-2.5.18}/src/pygpt_net/core/image/__init__.py

@@ -10,6 +10,8 @@
 # ================================================== #
 
 import os
+import uuid
+from time import strftime
 from typing import List
 
 from PySide6.QtCore import Slot, QObject
@@ -126,3 +128,16 @@ class Image(QObject):
             else:
                 return "_"
         return "".join(safe_char(c) for c in name).rstrip("_")[:30]
+
+    def gen_unique_path(self, ctx: CtxItem):
+        """
+        Generate unique image path based on context
+
+        :param ctx: CtxItem
+        :return: unique image path
+        """
+        img_id = uuid.uuid4()
+        dt_prefix = strftime("%Y%m%d_%H%M%S")
+        img_dir = self.window.core.config.get_user_dir("img")
+        filename = f"{dt_prefix}_{img_id}.png"
+        return os.path.join(img_dir, filename)

{pygpt_net-2.5.17 → pygpt_net-2.5.18}/src/pygpt_net/core/models/__init__.py

@@ -6,7 +6,7 @@
 # GitHub:  https://github.com/szczyglis-dev/py-gpt   #
 # MIT License                                        #
 # Created By  : Marcin Szczygliński                  #
-# Updated Date: 2025.06.24 02:00:00                  #
+# Updated Date: 2025.06.26 16:00:00                  #
 # ================================================== #
 
 import copy
@@ -385,18 +385,26 @@ class Models:
         :param mode: mode (initial)
         :return: mode (supported)
         """
-        if model.is_supported(MODE_CHAT):
-            self.window.core.debug.info(
-                "WARNING: Switching to chat mode (model not supported in: {})".format(mode))
-            mode = MODE_CHAT
-        elif model.is_supported(MODE_RESEARCH):
+        # if OpenAI API model and not llama_index mode, switch to Chat mode
+        if model.is_openai():
+            if model.is_supported(MODE_CHAT) and mode != MODE_LLAMA_INDEX:  # do not switch if llama_index mode!
+                self.window.core.debug.info(
+                    "WARNING: Switching to chat mode (model not supported in: {})".format(mode))
+                return MODE_CHAT
+
+        # Research / Perplexity
+        if model.is_supported(MODE_RESEARCH):
             self.window.core.debug.info(
                 "WARNING: Switching to research mode (model not supported in: {})".format(mode))
             mode = MODE_RESEARCH
+
+        # Llama Index / Chat with Files
         elif model.is_supported(MODE_LLAMA_INDEX):
             self.window.core.debug.info(
                 "WARNING: Switching to llama_index mode (model not supported in: {})".format(mode))
             mode = MODE_LLAMA_INDEX
+
+        # LangChain
         elif model.is_supported(MODE_LANGCHAIN):
             self.window.core.debug.info(
                 "WARNING: Switching to langchain mode (model not supported in: {})".format(mode))

{pygpt_net-2.5.17 → pygpt_net-2.5.18}/src/pygpt_net/core/models/ollama.py

@@ -6,10 +6,10 @@
 # GitHub:  https://github.com/szczyglis-dev/py-gpt   #
 # MIT License                                        #
 # Created By  : Marcin Szczygliński                  #
-# Updated Date: 2025.06.24 16:00:00                  #
+# Updated Date: 2025.06.26 16:00:00                  #
 # ================================================== #
-import os
 
+import os
 import requests
 
 class Ollama:
@@ -31,7 +31,8 @@ class Ollama:
         api_base = "http://localhost:11434"
         if 'OLLAMA_API_BASE' in os.environ:
             api_base = os.environ['OLLAMA_API_BASE']
-        print("Using Ollama base URL:", api_base)
+        self.window.core.idx.log("Using Ollama base URL: {}".format(api_base))
+
         url = api_base + "/api/tags"
         try:
             response = requests.get(url, timeout=2)

{pygpt_net-2.5.17 → pygpt_net-2.5.18}/src/pygpt_net/data/config/config.json

@@ -1,8 +1,8 @@
 {
   "__meta__": {
-    "version": "2.5.17",
-    "app.version": "2.5.17",
-    "updated_at": "2025-06-25T00:00:00"
+    "version": "2.5.18",
+    "app.version": "2.5.18",
+    "updated_at": "2025-06-26T00:00:00"
   },
   "access.audio.event.speech": false,
   "access.audio.event.speech.disabled": [],
@@ -308,6 +308,7 @@
   "render.plain": false,
   "render.code_syntax": "github-dark",
   "remote_tools.web_search": true,
+  "remote_tools.image": false,
   "send_clear": true,
   "send_mode": 2,
   "store_history": true,