Is your feature request related to a problem? Please describe.
When a content recognition result document is loaded, NVDA does not automatically read the content. The user must manually press NVDA+A (Say All) to hear the full result. This adds an extra, repetitive step for users whose primary goal is to listen to the entire recognized text from top to bottom.
Describe the solution you'd like
I propose adding a dedicated option to automatically perform a "Say All" command once a recognition result document has finished loading.
I believe the best place for this new option would be within the "Windows OCR" settings panel in NVDA's settings. It could be a checkbox labeled something like, "Automatically say all after recognition is complete."
When this option is enabled, any result document generated by a recognition feature (such as the built-in Windows OCR or any add-on using the contentRecog framework) would be read aloud automatically from the beginning, without requiring further user interaction.
Describe alternatives you've considered
An alternative I considered was to extend the functionality of the existing "Automatic Say All on page load" option (found in the 'Browse Mode' settings) to also cover recognition result documents.
However, I believe this would be confusing and undesirable. Many users, including myself, would only want this automatic reading behavior for recognition results, but not for web pages. Merging these two distinct contexts into a single option would force users to rely on complex workarounds, such as configuration profiles, to enable the feature only when needed. Therefore, a separate and specific option for recognition results is a much cleaner and more user-friendly solution.
Additional context
This feature request stems from my recent experience developing and using add-ons based on the contentRecog framework, which leverage LLMs for image description and OCR. In the vast majority of cases, the goal of running these tools is to listen to the entire output. Automating the "Say All" step would be a significant quality-of-life improvement and would make the workflow for using these recognition tools much more efficient.
Is your feature request related to a problem? Please describe.
When a content recognition result document is loaded, NVDA does not automatically read the content. The user must manually press NVDA+A (Say All) to hear the full result. This adds an extra, repetitive step for users whose primary goal is to listen to the entire recognized text from top to bottom.
Describe the solution you'd like
I propose adding a dedicated option to automatically perform a "Say All" command once a recognition result document has finished loading.
I believe the best place for this new option would be within the "Windows OCR" settings panel in NVDA's settings. It could be a checkbox labeled something like, "Automatically say all after recognition is complete."
When this option is enabled, any result document generated by a recognition feature (such as the built-in Windows OCR or any add-on using the contentRecog framework) would be read aloud automatically from the beginning, without requiring further user interaction.
Describe alternatives you've considered
An alternative I considered was to extend the functionality of the existing "Automatic Say All on page load" option (found in the 'Browse Mode' settings) to also cover recognition result documents.
However, I believe this would be confusing and undesirable. Many users, including myself, would only want this automatic reading behavior for recognition results, but not for web pages. Merging these two distinct contexts into a single option would force users to rely on complex workarounds, such as configuration profiles, to enable the feature only when needed. Therefore, a separate and specific option for recognition results is a much cleaner and more user-friendly solution.
Additional context
This feature request stems from my recent experience developing and using add-ons based on the contentRecog framework, which leverage LLMs for image description and OCR. In the vast majority of cases, the goal of running these tools is to listen to the entire output. Automating the "Say All" step would be a significant quality-of-life improvement and would make the workflow for using these recognition tools much more efficient.