Let me first thank everyone for your work on this, I have been using the old add-on for a very long time and this seems to be a huge improvement. My difficulty is that I can't use it at the moment. No matter whether I am using google or OpenAI, NVDA gets very sluggish, then the add-on gives me an error. Below is the section of the log I got with OpenAI. If anyone could let me know how best to troubleshoot, I would be most grateful. Whatever I enter, whether a conversational prompt or an image to describe, I get something like this. Note that this is nvda 2025.3.3 and Windows 10.
INFO - watchdog.waitForFreezeRecovery (05:07:34.490) - watchdog (11564):
Starting freeze recovery after 0.5310191000025952 seconds.
INFO - watchdog.waitForFreezeRecovery (05:07:36.084) - watchdog (11564):
Recovered from freeze after 2.125312500000291 seconds.
INFO - watchdog.waitForFreezeRecovery (05:07:36.624) - watchdog (11564):
Starting freeze recovery after 0.546387100002903 seconds.
INFO - watchdog.waitForFreezeRecovery (05:07:36.887) - watchdog (11564):
Recovered from freeze after 0.809432599999127 seconds.
INFO - watchdog.waitForFreezeRecovery (05:07:37.738) - watchdog (11564):
Starting freeze recovery after 0.5309941999948933 seconds.
INFO - watchdog.waitForFreezeRecovery (05:07:39.610) - watchdog (11564):
Recovered from freeze after 2.402896299994609 seconds.
INFO - watchdog.waitForFreezeRecovery (05:07:40.508) - watchdog (11564):
Starting freeze recovery after 0.5309695999967516 seconds.
INFO - watchdog.waitForFreezeRecovery (05:07:41.126) - watchdog (11564):
Recovered from freeze after 1.1497812999950838 seconds.
INFO - watchdog.waitForFreezeRecovery (05:07:51.213) - watchdog (11564):
Starting freeze recovery after 10.153553099997225 seconds.
INFO - watchdog.waitForFreezeRecovery (05:07:52.080) - watchdog (11564):
Recovered from freeze after 11.020148699994024 seconds.
INFO - watchdog.waitForFreezeRecovery (05:08:21.071) - watchdog (11564):
Starting freeze recovery after 0.5464904000036768 seconds.
INFO - watchdog.waitForFreezeRecovery (05:08:21.257) - watchdog (11564):
Recovered from freeze after 0.7321690000026138 seconds.
INFO - watchdog.waitForFreezeRecovery (05:09:44.457) - watchdog (11564):
Starting freeze recovery after 0.5154908999975305 seconds.
INFO - watchdog.waitForFreezeRecovery (05:09:50.165) - watchdog (11564):
Recovered from freeze after 6.2237103000006755 seconds.
INFO - watchdog.waitForFreezeRecovery (05:09:51.155) - watchdog (11564):
Starting freeze recovery after 0.5310664999997243 seconds.
INFO - watchdog.waitForFreezeRecovery (05:09:51.418) - watchdog (11564):
Recovered from freeze after 0.7939415000000736 seconds.
ERROR - external:globalPlugins.AIHub.chatcompletion.CompletionThread.run (05:09:59.654) - Thread-215 (3404):
Error processing response for model gpt-4.1:
Traceback (most recent call last):
File "C:\Users\test\AppData\Roaming\nvda\addons\OpenAI\globalPlugins\AIHub\chatcompletion.py", line 416, in run
self._responseWithStream(response, block, debug)
File "C:\Users\test\AppData\Roaming\nvda\addons\OpenAI\globalPlugins\AIHub\chatcompletion.py", line 512, in _responseWithStream
for event in response:
File "C:\Users\test\AppData\Roaming\nvda\addons\OpenAI\globalPlugins\AIHub\apiclient.py", line 856, in _stream_chat
content, think_from_tags = _apply_think_chain_to_chunk(content, think_chain_states)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "C:\Users\test\AppData\Roaming\nvda\addons\OpenAI\globalPlugins\AIHub\apiclient.py", line 612, in _apply_think_chain_to_chunk
out, frag = _split_think_pair_stream(out, states[idx], open_tag, close_tag)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "C:\Users\test\AppData\Roaming\nvda\addons\OpenAI\globalPlugins\AIHub\apiclient.py", line 566, in _split_think_pair_stream
resp, reasoning, in_think = _split_think_pair_inline(process, in_think=in_think, open_tag=open_tag, close_tag=close_tag)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "C:\Users\test\AppData\Roaming\nvda\addons\OpenAI\globalPlugins\AIHub\apiclient.py", line 525, in _split_think_pair_inline
resp_parts.append(text[i:j])
MemoryError
ERROR - external:globalPlugins.AIHub.conversation_dialog.ConversationDialog.OnResult (05:09:59.829) - MainThread (7832):
OpenAI add-on error:
NoneType: None
Let me first thank everyone for your work on this, I have been using the old add-on for a very long time and this seems to be a huge improvement. My difficulty is that I can't use it at the moment. No matter whether I am using google or OpenAI, NVDA gets very sluggish, then the add-on gives me an error. Below is the section of the log I got with OpenAI. If anyone could let me know how best to troubleshoot, I would be most grateful. Whatever I enter, whether a conversational prompt or an image to describe, I get something like this. Note that this is nvda 2025.3.3 and Windows 10.
INFO - watchdog.waitForFreezeRecovery (05:07:34.490) - watchdog (11564):
Starting freeze recovery after 0.5310191000025952 seconds.
INFO - watchdog.waitForFreezeRecovery (05:07:36.084) - watchdog (11564):
Recovered from freeze after 2.125312500000291 seconds.
INFO - watchdog.waitForFreezeRecovery (05:07:36.624) - watchdog (11564):
Starting freeze recovery after 0.546387100002903 seconds.
INFO - watchdog.waitForFreezeRecovery (05:07:36.887) - watchdog (11564):
Recovered from freeze after 0.809432599999127 seconds.
INFO - watchdog.waitForFreezeRecovery (05:07:37.738) - watchdog (11564):
Starting freeze recovery after 0.5309941999948933 seconds.
INFO - watchdog.waitForFreezeRecovery (05:07:39.610) - watchdog (11564):
Recovered from freeze after 2.402896299994609 seconds.
INFO - watchdog.waitForFreezeRecovery (05:07:40.508) - watchdog (11564):
Starting freeze recovery after 0.5309695999967516 seconds.
INFO - watchdog.waitForFreezeRecovery (05:07:41.126) - watchdog (11564):
Recovered from freeze after 1.1497812999950838 seconds.
INFO - watchdog.waitForFreezeRecovery (05:07:51.213) - watchdog (11564):
Starting freeze recovery after 10.153553099997225 seconds.
INFO - watchdog.waitForFreezeRecovery (05:07:52.080) - watchdog (11564):
Recovered from freeze after 11.020148699994024 seconds.
INFO - watchdog.waitForFreezeRecovery (05:08:21.071) - watchdog (11564):
Starting freeze recovery after 0.5464904000036768 seconds.
INFO - watchdog.waitForFreezeRecovery (05:08:21.257) - watchdog (11564):
Recovered from freeze after 0.7321690000026138 seconds.
INFO - watchdog.waitForFreezeRecovery (05:09:44.457) - watchdog (11564):
Starting freeze recovery after 0.5154908999975305 seconds.
INFO - watchdog.waitForFreezeRecovery (05:09:50.165) - watchdog (11564):
Recovered from freeze after 6.2237103000006755 seconds.
INFO - watchdog.waitForFreezeRecovery (05:09:51.155) - watchdog (11564):
Starting freeze recovery after 0.5310664999997243 seconds.
INFO - watchdog.waitForFreezeRecovery (05:09:51.418) - watchdog (11564):
Recovered from freeze after 0.7939415000000736 seconds.
ERROR - external:globalPlugins.AIHub.chatcompletion.CompletionThread.run (05:09:59.654) - Thread-215 (3404):
Error processing response for model gpt-4.1:
Traceback (most recent call last):
File "C:\Users\test\AppData\Roaming\nvda\addons\OpenAI\globalPlugins\AIHub\chatcompletion.py", line 416, in run
self._responseWithStream(response, block, debug)
File "C:\Users\test\AppData\Roaming\nvda\addons\OpenAI\globalPlugins\AIHub\chatcompletion.py", line 512, in _responseWithStream
for event in response:
File "C:\Users\test\AppData\Roaming\nvda\addons\OpenAI\globalPlugins\AIHub\apiclient.py", line 856, in _stream_chat
content, think_from_tags = _apply_think_chain_to_chunk(content, think_chain_states)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "C:\Users\test\AppData\Roaming\nvda\addons\OpenAI\globalPlugins\AIHub\apiclient.py", line 612, in _apply_think_chain_to_chunk
out, frag = _split_think_pair_stream(out, states[idx], open_tag, close_tag)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "C:\Users\test\AppData\Roaming\nvda\addons\OpenAI\globalPlugins\AIHub\apiclient.py", line 566, in _split_think_pair_stream
resp, reasoning, in_think = _split_think_pair_inline(process, in_think=in_think, open_tag=open_tag, close_tag=close_tag)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "C:\Users\test\AppData\Roaming\nvda\addons\OpenAI\globalPlugins\AIHub\apiclient.py", line 525, in _split_think_pair_inline
resp_parts.append(text[i:j])
MemoryError
ERROR - external:globalPlugins.AIHub.conversation_dialog.ConversationDialog.OnResult (05:09:59.829) - MainThread (7832):
OpenAI add-on error:
NoneType: None